Vital Dust: Life as a Cosmic Imperative [Paperback ed.] 0465090451, 9780465090457

Citation preview

■ »

Christian de Duve NOBEL

L A U R E AT E

Is the emergence of life on Earth the result of a single chance event or a combination of lucky accidents, or is it the inevitable outcome of biochemical forces woven into the fabric of the universe? And if inevitable, what are these forces, and how do they account not only for the origin of life but also for its evolution toward increasing complexity? Carefully forming analogies that present the mysteries of life in terms as familiar as a deck of cards or the letters of the alphabet, Christian de Duve, the Nobel Prize-winning biochemist, introduces readers to the most recent scien¬ tific theories about our ancient origins. This is a groundbreaking history of life on Earth, a history that only someone of de Duve’s stature and erudition could have written. The author guides us on a wondrous jour¬ ney through the past four billion years, from the formation of the first biomole¬ cules to the complexities of the human mind, from microscopic chains of amino acids and nucleotides to cataclysmic events in distant galaxies, arriving at the com¬ pelling conclusion that the universe is strewn with “vital dust” capable of spawn¬ ing life anywhere under the right condi¬ tions. Life and mmc are not accidents; they are natural manifestations of matter. At the heart of Vital Dust is the concept of seven increasingly complex “ages” of life on Earth. With each age, de Duve shows the key event that defined the age and the new event that led to the next. He argues that simple, deterministic chemical reac¬ tions put life on track but that other mecha¬ nisms led inexorably to greater complexity and biodiversity: the development of a lock-and-key system that serves as the uni¬ versal device of biological recognition at

BOSTON PUBLIC LIBRARY

Digitized by the Internet Archive in 2017 with funding from Kahle/Austin Foundation

https://archive.org/details/vitaldustlifeascOOdedu

VITAL DUST

VITAL DUST LIFE

AS

A

COSMIC IMPERATIVE

Christian de Duve

BasicBooks A Division of

HarperCol 1 i nsPublishers

Copyright © 1995 by the Christian Rene de Duve Trust. Published by BasicBooks, A Division of HarperCollins Publishers, Inc. All rights reserved. Printed in the United States of America. No part of this book may be reproduced in any manner whatsoever without written permission except in the case of brief quotations embodied in critical arti¬ cles and reviews. For information, address BasicBooks, 10 East 53rd Street, New York, NY 10022-5299. Designed by Ellen Levine Library of Congress Cataloging-in-Publication Data De Duve, Christian. Vital dust: life as a cosmic imperative / [Christian Rene de Duve]. p.

cm.

Includes bibliographical references and index. ISBN 0^165-09044-3 1. Life—Origin. 2. Life (Biology). 3. Evolution (Biology). I. Title. QH325.D42 1994 577—dc20

94-12964 CIP

95 96 97 98 ♦/RRD 987654321

To Life

Contents

List of Tables and Figures Preface Introduction The Unity of Life; The Tree of Life; The Antiquity of Life; The Cradle of Life; The Probability of Life; Foresight Excluded; The Ages of Life

PART I THE AGE OF CHEMISTRY Chapter 1

The Search for Origins The Setting; Chemists on the Trail; Searching the Skies; The First Steps; The Way to RNA; The Lesson of Metabolism

Chapter 2

The First Catalysts of Life Catalysis Without Proteins; The Case for Prebiotic Peptides; Introducing Thioesters; Catalytic Multimers to the Rescue

Chapter 3

The Fuel of Emerging Life The Problem of Primeval Membranes; The Case of the Miss¬ ing Hydrogen; The Case of the Excess Water; How the Wheels Are Kept Turning; The Thioester World

Chapter 4

The Advent of RNA The ATP Connection; Coenzymes: Children of the RNA World?; The Chemical Foundation of Life

CONTENTS

PART II THE AGE OF INFORMATION Chapter 5

RNA Takes Over

55

A Glimpse of Things to Come; The Magic Cipher; The Start of Replication; Darwin Plays with Molecules; The Birth of Proteins

Chapter 6

The Code

65

Darwin Needs Cells; Anatomy of Translation; The Origin of Translation; Structure of the Code; Metabolism Replaces Protometabol ism

Chapter 7

Genes in the Making

75

The Modular Game; RNA Splicing; The Advent ofDNA; Genetic Organization

Chapter 8

Freedoms and Constraints

83

Complementarity; Contingency; Modular Assembly

PART III THE AGE OF THE PROTOCELL Chapter 9

Encapsulating Life

89

The Timing of Cellularization; Cell Boundaries; Mechanisms of Cellularization; The Assembly of Membranes; The Con¬ struction of Outer Defenses; The Necessary Inlets and Out¬ lets; Cell Division

Chapter 10

Turning Membranes into Machines

99

Protonmotive Electron Transfer; The Attainment of Autonomy

Chapter 11

Adaptation to Life in Confinement

107

Sensing; Motility; Protein Export and the Birth of Digestion; A Touch of Sex

Chapter 12

The Ancestor of All Life

112

Reconstructing the Distant Past; Portrait of an Ancestor

Chapter 13

The Universality of Life Is Life Unique?; Extraterrestrial Life; Artificial Life

118

CONTENTS

IX

PART IV THE AGE OF THE SINGLE CELL Chapter 14

B acterfa Conquer the World

125

The Secret of Bacterial Success; The First Fork; Outlandish Colonies; The Green Revolution; The Great Oxygen Crisis

Chapter 15

The Making of a Eukaryote

137

' The Prokaryote-Eukaryote Transition; Giardia, a Living Fossil; The Eukaryotic Nucleus

Chapter 16

The Primitive Phagocyte

147

Spread of a Network; The Indispensable Props and Machines; Why Two Nuclei? Sex and the Single Cell

Chapter 17

The Guests That Stayed

160

An Age-Old Battle; Mitochondria: The CelTs Power Plants; Peroxisomes: Protectors Against Oxygen Toxicity; Chloroplasts: The Eukaryotic Link to the Sun; Other Possible Endosymbionts; A Last Look Back

PART V THE AGE OF MULTICELLULAR ORGANISMS Chapter 18

The Benefits of Cellular Collectivism

171

Slime Molds: An Instructive Example; The Importance of Sexual Reproduction; Principles of Cellular Collectivism

Chapter 19

The Greening of the Earth

176

Algae and Seaweeds; Mosses Invade the Lands; Vascularization, a Critical Acquisition; The Permian Crisis and the Formation of Seeds; Flowers and Fruits: The Crowning Achievement; Underground Infiltrators

Chapter 20

The First Animals Phytogeny and Ontogeny; The Awakening of Animal Life; The Worm’s Finest Hour; The “Milieu Interieur” and the Oxygen Connection

187

CONTENTS

X

Chapter 21

Animals Fill the Oceans

196

Body Duplication: The Road to Innovation; Invertebrates Galore; A Fateful Flip-Flop: From Mouth First to Mouth Second; The Birth of Vertebrates; The Echinoderms: An Evo¬ lutionary Quirk

Chapter 22

Animals Move Out of the Sea

203

Insects and Their Relatives: The Great Land Conquerors; Amphibians: The First Fish Out of Water; Reptiles “Invent ” the Amniotic Egg; The Mammalian Womb: The Ultimate Generation Machine; The Conquest of the Skies; The Driving Force of Evolution

Chapter 23

The Web of Life

214

The Primordial Link; Biospheric Metabolism; The Environment; Gaia

Chapter 24

The Virtues of Junk DNA

222

Selfish DNA; Split Genes; The Origin oflntrons; The Universe of Exons

PART VI THE AGE OF THE MIND Chapter 25

The Step to Human

229

Up in the Trees and Down Again; Mitochondrial Eve; Adam’s Apple

Chapter 26

The Brain

236

The Magic of Neurons; The Equipment of a Head; The Wiring of a Brain; On the Importance of Being Retarded

Chapter 27

The Workings of the Mind

245

Brain and Mind; Emergence of the Mind; Consciousness Explained?; The Rise and Fall of Dualism; Mind Power; The Self and Free Will

Chapter 28

The Works of the Mind Cultural Evolution; Culture and Mind; Artificial Intelligence

257

Values

Chapter 29

261

The Shaping of Human Societies; All Created Equal?; The Biology of Ethical Values 0

PART VII THE AGE OF THE UNKNOWN The Future of Life

Chapter 30 '

271

Natural Selection Derailed; Seven Heads, One Body; The Heart of the Problem; The Role of Science; The Next Five Billion Years

Chapter 31

The Meaning of Life

286

A Tale of Two Frenchmen; A Necessary Dialogue; The Living Cosmos; The Thinking Cosmos; A Meaningful Universe; Epi¬ logue

Notes

303

Glossary

325

Additional Reading

341

Index

351

Tables and Figures

Table INT. 1

The Seven Ages of Life on Earth

10

Figure 6.1

The Main Steps in Protein Synthesis

68

Figure 19.1

A Bird’s-Eye View of Plant Evolution

184

Figure 20.1

Some Key Steps in Early Animal Evolution

189

Figure 21.1

A Bird’s-Eye View of Invertebrate Evolution

202

Figure 22.1

A Bird’s-Eye View of Vertebrate Evolution

211

Figure 23.1

A Summarized History of Life on Earth

215

Figure 31.1

Two Views of the Tree of Life

298-99

Preface >>

It is enough for me to contemplate the mystery of conscious life perpetuating itself through all eternity, to reflect upon the marvelous structure of the universe which we can dimly per¬ ceive, and to try humbly to comprehend even an infinitesimal part of the intelligence manifested in nature.

—Albert Einstein

Plenty of books have been written on the origin of life, genes, cells, evolu¬

tion, biodiversity, the advent of humankind, brain, consciousness, society, the envi¬ ronment, the future of life, its meaning or lack of it. No one has had the temerity to handle all these subjects at the same time, for the simple reason that no one can master more than one or two of them, let alone all. Though no exception to such limitation, I have ventured beyond the boundaries of my competence because I feel that the attempt must be made if we are to understand the universe and our place in it. Life is the most complex phenomenon known to us, and we are the most com¬ plex beings so far produced by life. This book represents my attempt to look at the “bigger picture.” It goes back to a naive dream, conceived almost sixty years ago when, as a young medical student at the Catholic University of Louvain, Belgium, I first entered the field of science. What lured me into the laboratory, besides the fun of tackling problems, was the urge to understand. It seemed to me that science, by its insistence on rationality and objectivity, offered the best way to approach truth. The study of life looked particu¬ larly promising. It was going to be my pathway to the truth: per vivum ad verum. The dream soon receded. The demands of schooling and training—first in medi¬ cine, then in chemistry, finally in biochemistry—the struggle to establish a research group in postwar Belgium, the excitement of discoveries that led me to join the small band of investigators exploring living cells with modern methods, an appointment in 1962 that led me to share my time between my Belgian alma mater and the Rocke-

PREFACE

XIV

feller Institute (now the Rockefeller University) in New York, the duties and obliga¬ tions of academic life, the additional burden of founding a biomedical research insti¬ tute in Brussels, and, in the middle of it all, a disruptive trip to Stockholm in 1974, all conspired to keep me busy with day-to-day problems, leaving little time for wider issues. Active science narrows the mind more often than it broadens it, the reason being the increased specialization of facts, concepts, and techniques. As we dig deeper, our scope shrinks. An invitation to deliver the 1976 Alfred E. Mirsky Christmas Lectures at the Rockefeller University started to draw me out of my hole. The lectures were addressed to an audience of some 550 selected high school students from the New York area. I chose to “shrink” my young auditors a millionfold, equip them with appropriate “cytonaut” gear, and take them on a visit of the main sites to be found in a cell. By a combination of circumstances, a four-hour excursion became a fouryear expedition, which finally got into print in 1984 under the title of A Guided Tour of the Living Cell. To write and illustrate that book, I had first to turn into a

cytonaut myself, move outside my immediate research area, and explore parts of the cell with which I had only a passing acquaintance. It was an enjoyable experi¬ ence, the first step in a voyage of discovery that was to keep me busy for the fol¬ lowing ten years. The next step came when I began to reflect on the origin of the cells I had just toured. First their formation from primitive bacteria, a topic on which a number of revealing clues had been uncovered, leading back further to the origin of the first bacteria themselves. On this second question, I had always, like most biologists, uncritically accepted the standard version of a thickening soup of prebiotic chemi¬ cals somehow self-assembling into cells. I started looking more closely and soon found myself engrossed in the subject. It became my new research interest, result¬ ing in Blueprint for a Cell, published in 1991, which took a fresh look at the origin of life. This book ended with an affirmation—life is an obligatory manifestation of the combinatorial properties of matter—and with some questions: How about the further evolution of life? How about us? These questions defined the rest of my itinerary. The resulting trip, more hurried and sketchy than I would have liked—but time has become short—represents the nearest I shall ever get to accomplishing the dream of my youth. I offer this account with misgiving, aware of its inadequacies but hoping that it will stimulate others to further reflection. Even showing where I went wrong will be helpful. A warning: All through this book, I have tried to conform to the overriding rule that life be treated as a natural process, its origin, evolution, and manifestations, up to and including the human species, as governed by the same laws as nonliving processes. I exclude three “isms”: vitalism, which views living beings as made of matter animated by some vital spirit; finalism, or teleology, which assumes goaldirected causes in biological processes; and creationism, which invokes a literal acceptance of the biblical account. My approach demands that every step in the ori¬ gin and development of life on Earth be explained in terms of its antecedent and

PREFACE

xv

immediate physical-chemical causes, not of any outcome known to us today but hidden in the future at the time the events took place. Within this context. Vital Dust seeks to retrace the four-billion-year history of life on Earth, from the first biomolecules to the human mind and beyond. It takes the reader through seven successive “ages,” corresponding to seven levels of com¬ plexity: the Age of Chemistry, the Age of Information, the Age of the Protocell, the Age of the Single Cell, the Age of Multicellular Organisms, the Age of the Mind, and, challenging our insight, the Age of the Unknown, which includes the future and the timeless. The Age of Chemistry brings us straight to the essence of life, its universal aspect. Life, a chemical process, is to be understood in terms of chemistry. It started through the spontaneous formation and interaction of small organic molecules widely distributed in the universe. Given the physical-chemical conditions that pre¬ vailed on prebiotic Earth, these molecules were caught in a reaction spiral of grow¬ ing intricacy, eventually giving rise to the nucleic acids (RNA and DNA), proteins, and other complex molecules that dominate life today. This network of chemical reactions, formed almost four billion years ago, continues to provide the underpin¬ ning of all present-day manifestations of life. Although chemistry pervades this book, the reader will find no formula more complicated than H20 or C00. I focus on principles common to all forms of life on Earth. An important conclusion emerging from this consideration is that there must be congruence between protometabolism—the set of chemical reactions that first put life on track—and metabolism—the set of chemical reactions that support life today. Thus, our knowledge of present-day metabolism yields insights into life’s beginnings. Another lesson of the Age of Chemistry is that life is the product of determinis¬ tic forces. Life was bound to arise under the prevailing conditions, and it will arise similarly wherever and whenever the same conditions obtain. There is hardly any room for “lucky accidents” in the gradual, multistep process whereby life origi¬ nated. This conclusion is compellingly enforced when one considers the develop¬ ment of life as a chemical process. The Age of Information introduces molecular complementarity—the lock-andkey relationship—as a universal mechanism of biological recognition, which rules such diverse phenomena as enzyme specificity, self-assembly, communication among cells, immunity, hormonal effects, drug actions, and many other biological events. Its most fundamental manifestation is base pairing, the two-by-two joining of the main constituents of nucleic acids, first uncovered by Watson and Crick as the key to the double-helical structure of DNA, now known to govern all forms of genetic information transfer. In reviewing this crucial stage in the development of life, I emphasize underly¬ ing mechanisms. Base pairing arose from chemical events that had nothing to do with information transfer. Molecular replication, the offshoot of base pairing, was a fringe benefit of prebiotic chemistry. Once it emerged, however, replication opened

PREFACE

XVI

the way to hereditary continuity—based on accurate copying of genetic mes¬ sages—and to evolution—by way of mutations of the messages and screening by natural selection. But for this to happen, a machinery had to be put together for expressing the messages in a form suitable for natural selection to act on. Every step in the construction of this machinery was the product of deterministic chemical processes, modulated by selection. A key factor that first came into play with the Age of Information is contin¬ gency. Mutations are chance events, which fact, it is often claimed, implies a view of evolution as being ruled by chance. While not denying the role of contingency in evolution, I point out that chance operates within constraints—physical, chemi¬ cal, biological, environmental—that limit its free play. This notion of constrained contingency runs as a leitmotiv throughout my reconstruction of the history of life on Earth. The Age of the Protocell was a long period during which the main attributes of cellular organization were progressively assembled. Its outcome was an organism ancestral to all forms of life present on Earth today. The contention that all living organisms are derived from a common ancestor rests on overwhelming evidence. This organism emerged around 3.8 to 3.7 billion years ago. The Age of the Single Cell was dominated by two events. One was the evolution and diversification of bacteria, or prokaryotes, which now occupy almost every available niche on our planet. A fateful occurrence in this evolution was the appear¬ ance of organisms capable of using the energy of sunlight to extract from water the hydrogen needed for self-construction, thereby releasing molecular oxygen. This event is responsible for the rise in atmospheric oxygen that occurred between 2.0 and 1.5 billion years ago. It posed a major threat to the anaerobic forms of life that occupied Earth at that time. Exposed to increasing amounts of the, for them, toxic oxygen molecule, organisms had to adapt or perish. Many bacterial species suc¬ cumbed to the “oxygen holocaust”; those that survived did so with innovations that played a crucial role in further evolution. The second key event during the Age of the Single Cell was the prokaryoteeukaryote transition, the transformation of an ancestral bacterial cell into the much larger, more complex cells that make up algae, amoebae, yeasts, and many other unicellular organisms, as well as all plants, fungi, and animals, including humans. This epochal transformation, which may have taken as long as one billion years, led to the development of a primitive phagocyte (eating cell), a large, highly organized cell able to engulf and digest bacteria and other bulky objects. Cells of this type occasionally established mutually advantageous relationships with engulfed bacte¬ ria, which were retained as permanent guests, or endosymbionts, and evolved into functional cell parts, including mitochondria and chloroplasts. The need to adapt to oxygen may have precipitated this evolution. With the Age of Multicellular Organisms, life entered the phase most familiar to us. The Earth, which for some three billion years had harbored only invisible microorganisms, became progressively occupied, first in the waters and later on

PREFACE

XVII

land, by a gamut of increasingly complex plants and animals. This evolution was marked by successive improvements in reproductive strategy adapted to changing environments. A major step was the development of sexual reproduction. In the plant world, the progress of this development was from spores to seeds to flowers and fruits. In the animal world, haphazard aqueous fertilization gave way to copula¬ tion; fertilized egg cells were first laid and allowed to develop in water, then on land, in the sheltered confines of an amniotic egg, and finally inside a womb—for a short developmental stage in marsupials and, later, for a longer one in placentals. This evolution seems dominated by biodiversity, a profusion of species, prod¬ ucts of chance mutations that happened to confer an advantage in a particular envi¬ ronment. Within this variability, however, there is a trend toward complexification. The two features explain the structure of the “tree of life.” First, there is the trunk, shaped by a series of “fork organisms,” each affected by a mutation that signifi¬ cantly changed the body plan in the direction of greater complexity. Then there is the system of increasingly ramified branches, expressing increasingly trivial alter¬ ations of established body plans, the main source of diversity within each major group. This distinction reconciles two views of life that have often been opposed to one another in the past; it puts chance and necessity in correct perspective. Also important in the development of this tree was the expanding web of interrelation¬ ships linking living organisms with each other and with the environment in increas¬ ingly complex ecosystems. A concomitant of animal evolution was development of a brain. Once neurons appeared—which they did very early—they joined into increasingly elaborate net¬ works, driven at each step by the resulting evolutionary advantages. Out of the brain arose consciousness, inaugurating, in a manner that defies comprehension, the Age of the Mind. The latest stages in this evolution have been amazingly rapid, leading in only a few million years to the conversion of primates into humans. This event has dramatically modified the history of life on Earth, largely substi¬ tuting the fast, human-directed process of cultural evolution for the slow process of Darwinian evolution by natural selection. Art, science, philosophy, ethics, religion are products of this new age. So are medicine and technology, which have trans¬ formed the face of the Earth in the space of a few centuries, creating immense prob¬ lems that urgently challenge human ingenuity and wisdom. If we do not, in the near future, deal satisfactorily with these problems, especially the demographic explo¬ sion, which is at the root of most of them, natural selection will do it for us, but with consequences that may be tragic for humankind and for much of the living world. Such is the message we receive when we use our understanding of the his¬ tory of life to peer into the Age of the Unknown. Whatever happens, life will recover, as it has so many times in the past after major planetary catastrophes. Most likely, it will continue to evolve toward greater complexity. There is no reason why we should view ourselves as the pinnacle of a process that still has another five billion years to go. What form the next step will take, where and how it will happen, even what extant species will be involved, are

xviii

PREFACE

unanswerable questions. What will be recognized tomorrow as a fork organism is a mere terminal twig on the tree of life today. In the last chapter, I try to put it all together. From the perspective of determin¬ ism and constrained contingency that pervades the history of life as I have recon¬ structed it, life and mind emerge not as the results of freakish accidents, but as nat¬ ural manifestations of matter, written into the fabric of the universe. I view this universe not as a “cosmic joke,” but as a meaningful entity—made in such a way as to generate life and mind, bound to give birth to thinking beings able to discern truth, apprehend beauty, feel love, yearn after goodness, define evil, experience mystery. I make no explicit mention of God because this term is loaded with multi¬ ple interpretations linked to a variety of creeds. As a scientist, I have chosen to pro¬ vide a summary of available evidence and to share my personal interpretation of this evidence, leaving it to readers to draw their own conclusions. Lest I be misun¬ derstood, let me stress once more that the key word is chemistry, not some precon¬ ceived notion of how things ought to be. To whom is this book addressed? To everyone. The topic, which includes our nature, origin, history, and place in the universe, is of interest to all of us. Faced with a number of burning issues affecting the future of life on Earth, perhaps even the survival of humankind, it is imperative that we contemplate these problems within their natural context. We must learn to “think biologically” and act accord¬ ingly. Like most history books, Vital Dust includes parts that may interest some read¬ ers more than others. Although a thread of continuity runs through all seven parts, each has been written in such a way as to encourage browsing. This book is very much the product of personal reading and reflection. I owe an immense debt of gratitude to the many authors who have helped me by their thoughtful, well-documented, and enlightening expositions of their fields. I have done my best to give them due credit in the notes and references cited at the end of the book. I have also benefited greatly from conversations and discussions with a number of colleagues and friends. Thankfully mentioning their names in no way implies that they are ready to vouch for my presentation of scientific facts, even less that they approve my interpretations or share my ideas. I have in mind my long-time associate, friend, and present “boss,” Miklos Muller, who has greatly helped me with his encyclopedic knowledge of microorganisms; my newly made friends in the origin-of-life field, including Gustaf Arrhenius, Manfred Eigen, Albert Eschenmoser, Stanley Miller, Leslie Orgel, William Schopf, Arthur Weber, and many oth¬ ers; Stuart Kauffman, who has introduced me to the intricacies of “artificial life”; and Francis Crick and Gerald Edelman, who have tried their best—largely in vain, I regret to say—to convert me to thinking correctly about the brain. My son Thierry deserves my thanks for guiding me through the intricacies of Kant.

PREFACE

xix

My greatest debt is to my former publisher and editor, my faithful friend Neil Patterson, who has sacrificed endless hours of his valuable time to put this book in acceptable shape. He has not only trimmed flowery adjectives, chatty asides, irrele¬ vant remarks, ponderous constructions, and other infelicities. He has drawn my attention to a number of mistakes and obscurities, clipped some of my more exu¬ berant or incautious statements. My thanks extend to Ippy Patterson for so beauti¬ fully drawing the tree of life, something of a symbol of this book. My gratitude goes also to my publishers at Basic Books, especially Susan Rabiner, who has made a number of valuable suggestions concerning the organiza¬ tion of the book, and to Suzanne Wagner, the copyeditor, and Michael Mueller, managing editor, who have been most helpful in putting the book in its final shape. Finally, I wish to thank my children, Thierry, Anne, Frangoise, and Alain, for the word processor they gave me on the occasion of my seventieth birthday. This frightening gift—I had never even used a typewriter in my life—has become a trusted and valued assistant. Its use has not, however, diminished my constant calls on the services of two highly competent and devoted “flesh and blood” assistants, Anna Polowetzky (Karrie), in New York, and Monique Van de Maele, in Brussels. Only my wife, Janine, can tell how much she has endured while I was struggling with my unwieldy project. I thank her with my love. Nethen and New York January 31, 1994

VITAL DUST

Introduction

This book is about the history of life on Earth—from its birth, shrouded in

the depths of the past, to the variegated pageantry of living beings that cover our planet today. It is the most extraordinary adventure in the known universe, an adventure that has produced a species capable of influencing in decisive fashion the future unfolding of the natural process by which it was born. The history of life is marked by a series of innovations, each introducing a new level of complexity, each to be accounted for in terms of the natural laws of physics and chemistry. Before we embark on this voyage of discovery, I shall define a few general notions that will be with us all along the way.

THE UNITY OF LIFE Life is one. This fact, implicitly recognized by the use of a single word to encom¬ pass objects as different as trees, mushrooms, fish, and humans, has now been established beyond doubt. Each advance in the resolving power of our tools, from the hesitant beginnings of microscopy little more than three centuries ago to the incisive techniques of molecular biology, has further strengthened the view that all extant living organisms are constructed of the same materials, function according to the same principles, and, indeed, are actually related. All are descendants of a sin¬ gle ancestral form of life. This fact is now established thanks to the comparative sequencing of proteins and nucleic acids. These two groups of substances, which are the most important constituents of all forms of life, are entirely different chemically but are both long chains made by the stringing together of a large number of molecular units—up to several hundred for the proteins, often considerably more for the nucleic acids. Think of strings of beads of different colors, of trains made of different kinds of cars hooked end to end, or, more appropriately, of very long words assembled with

9

VITAL DUST

different letters. The beads, cars, or letters that make up proteins are called amino acids; those that form nucleic acids are called nucleotides. Protein “words” are made with an alphabet of twenty kinds of amino-acid letters; nucleic-acid “words” with an alphabet of four kinds of nucleotide letters. High-performance methods now exist for establishing the exact order in which the building blocks of these natural macromolecules follow each other in a given chain. These techniques allow scientists to decipher with great accuracy the sequences of amino acids in proteins and of nucleotides in nucleic acids, the “spelling” of molecular words. The fine print in the book of life is now legible. One fact of immense importance has emerged from this newly gained molecular literacy. Organisms as different as a microbe, a corn plant, a butterfly, and a human being contain similar proteins and nucleic acids. These similarities are much closer than could possibly be accounted for on the basis of chance. They enforce the in¬ escapable conclusion that these molecules and, therefore, all organisms throughout the living world are related to each other, derived from a common ancestor. As a sim¬ ple analogy, compare the English word assembly with the French word assemblee used to convey the same notion. The words obviously did not arise independently in the two languages; they are related through a common ancestral word from which both are derived. The words are not identical because they have changed differently since they diverged from their common ancestor. The same is true for related macro¬ molecules. Their sequences differ because they have suffered changes transmissible from generation to generation—that is, mutations—as the various organisms that contain them have evolved after diverging from a common ancestor. This unity within diversity simplifies our task. We are trying to trace the history of life, not of lives. The common ancestor divides our itinerary into two parts. First, we must reconstruct the manner in which the common ancestor emerged from whatever materials were available before life appeared. Next, we must find out how all extant living organisms evolved from the common ancestor.

THE TREE OF LIFE It is common knowledge that life has left fossil traces of its history. Patient decrypt¬ ing of these vestiges has allowed paleontologists to conjure up from the remote past the ghosts of ancient plants and animals and to piece together a rough evolutionary history of the organisms that inhabit our planet today. However, the fossil record is very incomplete. Quite often, a single bone or tooth, the imprint of a leaf, or the hollow cast of a worm is all that is available to reconstruct an entire organism. In addition, the fossil record hardly goes back further than 600 million years. The ear¬ lier record is extremely sparse. Countless organisms must have existed that have left no trace or whose traces have not yet been unearthed. Were only fossil docu¬ ments available, however numerous and well preserved, a complete history of life

INTRODUCTION

3

could not be written or even contemplated. Our information comes not so much from dead remains as from living beings. The whole history of life is written into present-day organisms. All we need in order to reconstruct this history is to be able to read that text. This we can now do using the comparative sequencing of related macromole¬ cules from different kinds of organisms. Such analysis can serve to evaluate the evolutionary distance between a pair of organisms—sisters, first cousins, ten-timesremoved cousins, and so on—taking as a yardstick the number of differences between the compared sequences. The more numerous these differences—so the assumption goes, subject to a considerable number of caveats and qualifications— the longer thd time the molecules have been evolving separately, that is, the longer the time since the species possessing the molecules diverged from their last com¬ mon ancestor. With enough information of this kind, it is possible, in principle, to reconstruct the entire tree of life on the basis of the properties of organisms living around us today. In the linguistic analogy, such an approach amounts to molecular etymology. Imagine a linguist confronted with only contemporary texts in French, Italian, Spanish, and Romanian. Even without knowledge of the past, such an expert would be able to conclude from the many similarities among words of the same meaning that these four languages are related. By careful comparative studies based on the assumption that words change only gradually with time, he might even succeed in reconstructing the ancestral Latin, as well as the manner in which the four lan¬ guages evolved from it. This reconstruction would be shaky at first, easily led astray by coincidental similarities, appropriations by one language from another, and other red herrings. But it would become increasingly secure as more words were examined, analyzed, and compared. One of the earliest examples of molecular etymology—now a classic—concerns a protein called cytochrome c.1 This small protein, about one hundred amino acids long, participates in the utilization of oxygen by many living beings. The human ver¬ sion of cytochrome c differs from that of the rhesus monkey by a single amino acid and from those of the dog, rattlesnake, bullfrog, tuna fish, silkworm, wheat, and yeast by 11, 14, 18, 21, 31, 43, and 45 amino acids, respectively. Such figures provide es¬ timates of the increasingly remote times when each of these species branched from the last ancestor they have in common with us. These estimates, which fit with the fossil record for the animals, go further back in time, to kinships for which no fossil evidence exists. Note that even the wheat and yeast cytochrome c molecules share more than fifty amino acids with each other and with the human molecule, indis¬ putable proof that these three widely different species have a common ancestry. Comparative sequencing of cytochrome c was done more than twenty years ago. Since then, many proteins, as well as nucleic acids, have been similarly compared, and more are being analyzed every day. Interpretation of the data is not simple. Nevertheless, even though many uncertainties and controversies are left to be resolved, we are beginning to know in some detail and with some degree of confi-

4

VITAL DUST

dence the manner in which extant living forms originated from their common ancestor by a progressive branching of the tree of life. In its upper part, the molecu¬ lar tree agrees with that drawn by paleontologists on the basis of fossil evidence, except for a number of details that have been added or corrected with the help of the new data. The lower part of the tree is new to us. It has proved surprising.

THE ANTIQUITY OF LIFE The shape of the tree is gradually becoming clear. But what of its time scale? In paleontological work, the time coordinate is provided by the extensive geological and geochemical investigations that allow the age of a given rock formation to be estimated. If a fossil is found in a terrain believed by geologists to be 200 million years old, we know that the organism that left the remains lived 200 million years ago, give or take a few million years. For molecular trees, the unit of measure is not time but number of mutations: the changes, transmissible from generation to gener¬ ation, that the molecules have undergone in the course of evolution. Or, more accu¬ rately, the number of mutations compatible with survival and proliferation (the “tol¬ erable’' mutations), since other changes are eliminated by natural selection2 and leave no trace in extant molecules. In order to convert this unit into units of time, we must know the tolerable mutation rate. The time scale of a given tree will be very different if tolerable mutations are taken to occur, on average, every one mil¬ lion, two million, or ten million years. This is one of the major uncertainties of the molecular method. The best way to resolve the problem is to compare the molecu¬ lar trees with the paleontological trees. This works for the upper part of the tree, for which paleontological data are available. But what about the lower part? The answer has come, just in recent decades, from bacterial fossils. Bacteria are small entities, usually no more than a few hundred-thousandths of an inch in size, with shapes ranging from globular to threadlike, visible only with a good microscope. There are plenty of bacteria in the world today. For most of us, the name “bacteria” raises specters of plague, cholera, tuberculosis, leprosy, diph¬ theria, and other dreaded ills. However, disease-causing microbes are only a small minority among a wide diversity of harmless or useful forms, which occupy almost every possible kind of habitat, from the balmy shelter of the human gut to the brine of drying seas and the boiling waters of volcanic springs. The richest source of bac¬ teria is the soil, where these invisible organisms accomplish the all-important decomposition of dead plants and animals, recycling the constituents of life. Bacteria are the simplest forms of life and, as we long suspected and now know, the earliest ones. Fossil traces of these organisms would therefore be invaluable in reconstructing and timing the lower part of the tree of life. Such traces have been tound in the last decades.3 They are of two very different dimensions. At the visible level, the evidence comes from special layered rocks called stromatolites. Such for-

INTRODUCTION

5

mations are derived by fossilization from huge bacterial colonies composed of superimposed mats, each consisting of a different kind of bacterial species. The top layers of the colony are made of organisms, termed “phototrophic,” that use the energy of sunlight to make their own constituents; later, after they die, their sub¬ stance provides food for the underlying layers. Colonies of this sort cover large areas in certain coastal regions, for example, in Baja California in northwest Mex¬ ico. With time, the colonies fossilize into stromatolites by a process of which every stage is known from certain representative rocks. Stromatolites have been found in diverse terrains, in many different parts of the world, spanning all geological ages. Some date back to as much as 3.5 billion years ago, which, for all practical purposes, represents the limit of the useful geological record. It is possible that stroma¬ tolite-generating colonies existed even before that time, but their traces could not have survived geological transformations. The second type of evidence of early bacterial life is microscopic. Most bacteria are enclosed within a solid shell, or wall. This has allowed ancient bacteria to leave their imprints in mud, which later solidified into rocks, just as long-extinct ferns have left their delicate tracings, except that elaborate techniques and a healthy dose of critical discrimination are needed to identify a genuine bacterial microfossil and distinguish it from spurious traces and recent contaminants. A number of authentic imprints are now known. Interestingly, these traces are often found in stromatolites, providing additional proof, if any were needed, of the bacterial origin of these rocks. A few microfossils also date as far back as 3.5 billion years. So life is at least 3.5 billion years old. Such is the startling message delivered by stromatolites and microfossils. Compare this age with the limit of about 600 million years beyond which virtually no trace of plant or animal has been found and you can appreciate the enormous size of the hidden lower part of the tree of life: four to five times the size of the upper part, which encompasses the entire evolutionary history of plants and animals. During the immensely long time, almost three billion years, that preceded the emergence of the first plants and animals known to us by their fossils, life seems to have remained almost at a standstill. Stromatolites and microfossils do not look very different whether one billion or three billion years old. This appearance of stagnation is misleading, though. Events of cardinal impor¬ tance took place in the shadow of the stromatolites, preparing the great explosion of life forms that burst forth around 600 million years ago. According to their fossil traces, the bacteria that lived 3.5 billion years ago were diverse and advanced. They may even have included representatives of the most elaborate forms of phototrophic organisms known today. No doubt, these early life forms were preceded by more rudimentary ones, themselves pieceded by the com¬ mon ancestor of all life. When did this ancestral organism arise? Perhaps as early as 3.8 billion years ago, as suggested by physical analyses of fossil carbon deposits (kerogen) dating back to that time. These deposits show an enrichment in carbon atoms with atomic mass 12 (that is, with a mass equal to 12 times the mass ot the hydrogen atom) over carbon atoms with atomic mass 13. Such enrichment of the

VITAL DUST

6

lighter over the heavier carbon isotope4 is a characteristic feature of biological car¬ bon assimilation. An upper limit of four billion years ago for the earliest life form is set by the conditions that probably obtained on Earth in the beginning of its history. Experts tell us that the Earth first condensed from a cloud of gas and dust some 4.5 billion years ago. For the next 500 million years, the young planet, battered by falling asteroids and racked by violent volcanic eruptions, remained unfit for life.5 The common ancestor of life probably appeared on Earth some time between 4.0 and 3.8 billion years ago. Recognizing that these dates are uncertain, we shall adopt them as the most reasonable estimates available from the present state of our knowledge.

THE CRADLE OF LIFE Where did life originate? The obvious answer to this question—that life originated on Earth—is not accepted by everyone, partly for reasons of time. It appears from the evidence just discussed that a maximum of 200 million years may have been available for the emergence of the common ancestor of life from the materials our lifeless planet could offer. Although short with respect to the whole history of life on Earth, this still amounts to a very long time span in absolute terms. If we repre¬ sent the whole Christian era—two thousand years—by one inch, the time available for the emergence of life could measure as much as 1.5 miles. Yet there are some who see this time as too short for the development of something as complex as a bacterial cell. This attitude goes back to an early belief, no longer shared by most scientists, that life originated through an exceedingly slow and long process, per¬ haps too slow and too long for our planet to harbor. This belief is one reason why the suggestion has been made that life came to Earth from outer space. The possibility that life may have an extraterrestrial origin has been repeatedly considered.6 The theory was proposed at the turn of the century and defended with almost mystical fervor by Svante Arrhenius, a Swedish Nobel Prize-winning chemist who coined the term “panspermia” to express his belief that seeds of life exist everywhere in space and are showered continually on the Earth. More recently, a modified version of this theory has been advocated with equal forceful¬ ness by Fred Hoyle, a celebrated British astronomer, and his colleague Chandra Wickramasinghe, an astronomer from Sri Lanka, who have claimed that viruses and bacteria continually arise on the tails of comets and fall on the Earth with parti¬ cles of cometary dust.7 Some of these germs could be pathogenic and start epi¬ demics, which, according to these two scientists, may have played important roles in shaping human history. The human nose, they even speculated, may have devel¬ oped as an evolutionary protection against diseases caused by inhaling raindrops contaminated by extraterrestrial germs. Another proposal has been made, under the name “directed panspermia,” by Francis Crick, of double-helix fame, and Leslie

INTRODUCTION

7

Orgel, a pioneer in prebiotic chemistry. The two British-born American scientists, now at the Salk Institute for Biological Studies in La Jolla, California, have sug¬ gested that the first germs of life reached the Earth by a spaceship sent by some dis¬ tant civilization.8 With such distinguished proponents, panspermia can hardly be dismissed with¬ out a hearing. Critics of the theory have objected that living organisms could not withstand the intense radiation to which they would be exposed in outer space. But this claim has been disputed. Supporters of the theory have stated that life could not have originated on Earth for lack of time. On what basis they estimate 200 million years to be insufficient for the development of life is not clear. The real question is >*

whether there is solid evidence on which to base a surmise. There is none for a spaceship or for its senders. Things are different for comets and other celestial objects, such as meteorites. These bodies do contain organic molecules of the kind found in living organisms. In the opinion of most investigators, however, these sub¬ stances are produced by simple chemical reactions that take place “out there.” They are not made by living organisms. Of the existence of such organisms, there is as yet no convincing, or even suggestive, sign. Fairness demands that the matter be left open until the controversy is settled. But common sense and economy counsel that it be ignored in our further discussions. The best reason for doing so is that, even if we accept that life came to Earth from outer space, we are still left with the problem of how it originated. I shall, therefore, assume that life was bom right where it actually is: here on Earth.

THE PROBABILITY OF LIFE How did life originate? Would it emerge if we could move back in time and let events unfold in the same setting, or if the setting were duplicated on some other planet? If so, would it be the kind of life we know or something different? These are questions science has so far failed to answer. What we have instead is a profu¬ sion of theories, slanted by the scientific specializations, philosophical attitudes, and ideological biases of their authors. Two schools even go so far as to claim that the origin of life is not a valid problem for science to investigate. They do so for very different reasons, although both their reasons are rooted in the belief that life is an extremely improbable phenomenon. So improbable, according to the creation¬ ists, that nothing short of direct divine intervention can explain the emergence of even the simplest of living organisms. The more rationalist believers in the improb¬ ability of life reject this claim, pointing out that chance produces extremely improb¬ able events all the time. However, for the very reason they are improbable products of chance, such events are unique and nonreproducible, and therefore inaccessible to scientific investigation. To explain this point of view, I shall take an example from the game of bridge.

VITAL DUST

8

Bridge is a card game played by four players with a deck of fifty-two cards, including thirteen each of spades, hearts, diamonds, and clubs. The cards are shuf¬ fled and dealt around the table one by one. Suppose you are one of the players and you pick up all thirteen spades. Without doubt, you would speak of a fantastic stroke of luck. You would be right. The odds of being dealt all thirteen spades are one in 635 billion. Let armies of bridge players play day and night for centuries and the thirteen spades may never turn up even once. Indeed, to my knowledge, such an event has never been recorded in the annals of bridge.9 As the first recipient of this astonishing gift of chance, you would achieve instant world fame. Your name would appear in every bridge column and book. All very true and understandable, except that any other bridge hand has exactly the same probability of being dealt— one in 635 billion. Most hands, however, are not sufficiently spectacular to make history. Note that I have not taken into account in my estimate the cards the other players are getting. If the complete distribution of the cards is to be specified, the probabil¬ ity is of the order of one in fifty billion billion billion (5 x 1028). If all the human beings that ever existed had done nothing but play bridge day and night during their whole lives, the odds that a distribution dealt this evening at your bridge club had ever occurred before would still be very low. Yet in no bridge club do players exclaim at being witness to an extraordinarily improbable event every time the cards are dealt. This example illustrates the simple fact, not always recognized, that single events of very low probability take place all the time without anybody paying any notice unless there is something special about the event. The emergence of life, it has been said, could have been such an event, a fantastic stroke of luck, like getting thirteen spades at bridge, but no transgression of the laws of probability. If this were so, we would be wasting our time trying to explain the origin of life in scientific terms. A number of eminent scholars have made this claim. Some have even pushed it to its logical conclusion, that if life is a highly improbable product of chance, it has no place in any sort of cosmological view we may entertain. Let bil¬ lions of planets go through the same history as that of the Earth. Let even billions of big bangs give rise to billions of universes similar to ours. Nowhere would there be life. Its emergence was a lusus naturae, a cosmic joke. In the words of the late Jacques Monod, one of the greatest French biologists: “The universe was not preg¬ nant with life.”10 This statement has profound philosophical implications. These I shall address later. For the time being, I wish merely to examine the scientific validity of the probability argument. Its logic is impeccable, provided we are dealing with a single event. But the emergence of life cannot possibly have happened as a single event. To illustrate this impossibility, Hoyle has used the analogy of a Boeing 747 arising ready to fly from a tornado-swept junkyard.11 The possibility of a living cell com¬ ing together in one shot is immeasurably less plausible than the spontaneous assem¬ bly of a Boeing 747—if degrees of impossibility are to be envisaged. Only instant

INTRODUCTION

9

creation—a miracle—could accomplish such feats, and miracles, by definition, fall outside the boundaries of scientific investigation. They are a last recourse, when all attempts at a rational explanation have failed—a point, incidentally, that is difficult to identify, since the explanation may have to await new knowledge, as has so often been the case in times past. But we are far from having reached such a point with respect to the origin of life. The field is burgeoning with an abundance—almost a surfeit—of informative facts and stimulating ideas. A Boeing 747 is built piecemeal in a very large number of steps. Raw materials are first refined or synthesized and worked into a multitude of separate parts. These are then joined, in modular fashion, to make the engines, the body and wings, the flaps, the landing gear, the electronic circuits, and all the other parts of the aircraft. These various parts are then brought together for final assembly. The steps in the construction of a living cell are different, but the principle is the same. Because of the high complexity of the final product, there must, by necessity, be a very large number of steps, often modular in nature. This consideration completely alters the probability assessment. We are being dealt thirteen spades not once but thousands of times in succession! This is utterly impossible, unless the deck is doctored. What this doctoring implies with respect to the assembly of the first cell is that most of the steps involved must have had a very high likelihood of taking place under the prevailing conditions. Make them even moderately improbable and the process must abort, however many times it is initi¬ ated, because of the very number of successive steps involved. In other words, con¬ trary to Monod’s affirmation, the universe was—and presumably still is—pregnant with life. To me, this conclusion is inescapable. It is based on logic, not on an a priori philo¬ sophical tenet. It does not, however, imply that the emergence of life followed a rigid, preordained course. Even less does it mean that only one kind of life was or is possible. There is room in a deterministic pathway for bifurcations, alternative routes, accidents, even chaos, just as there are many ways for rainwater to run down a mountain. What counts are the constraints of the terrain. A smooth top can lead in many directions. Even a pebble can alter the course of a rivulet. On the other hand, a crater leading into a gorge will force water to flow in a single direction.

FORESIGHT EXCLUDED In the making of a Boeing 747, all steps are intentional, designed and organized according to a detailed blueprint of the final objective. Things cannot have been the same in the making of the first living cell. Every step had to stand on its own and cannot be viewed as preparation for things to come. This kind of objectivity is diffi¬ cult to sustain because we know the outcome and also because our whole thinking about life is permeated by intentionality. Cells are so obviously programmed to

VITAL DUST

10

develop according to certain lines, organs adapted to perform certain functions, organisms suited to certain environments, that the word design almost unavoidably comes to mind. A whole school of thought has been inspired by these appearances of design, maintaining that living organisms are actuated by final causes, in the Aristotelian sense of the term. Called finalism, this doctrine is close to vitalism, the belief that living organisms are animated by a vital principle. Both views are now largely discredited. Design has given place to natural selection. The vital principle has joined ether and phlogiston in the cemetery of discarded concepts. Life is increasingly explained strictly in terms of the laws of physics and chemistry. Its ori¬ gin must be accounted for in similar terms.

THE AGES OF LIFE History is a continuous process that we divide, in retrospect, into ages—the Stone Age, the Bronze Age, the Iron Age—each characterized by a major innovation added to previous accomplishments. This is true also of the history of life, which, so far, has gone through six major ascending planes of complexity (see table INT.l). First, there is the Age of Chemistry. It covers the formation of a number of major constituents of life, up to the first nucleic acids, and is ruled entirely by the univer¬ sal principles that govern the behavior of atoms and molecules. Then comes the Age of Information, thanks to the development of special infor¬ mation-bearing molecules that inaugurated the new processes of Darwinian evolu¬ tion and natural selection particular to the living world.

TABLE INT. 1 The Seven Ages of Life on Earth

Age

Millions of Years

The Eirth of Earth

4,550 before present

Chemistry 4,000-3,000 before present

Information The Protocell

_

The Single Cell

3,000-3,700 before present

Multicellular Organisms

700 - 600 before present

The Mind

6 before present

The Unknown

present

The End of Earth

5,000 after present

INTRODUCTION

11

The third stage in the history of life is the Age of the Protocell, the first living unit surrounded by a membrane and capable of acquiring a number of key proper¬ ties linked to this feature. This age ends with the emergence of the common ances¬ tor of all life on Earth. Next comes the Age of the Single Cell, spanning more than two billion years divided into two major phases, the prokaryotic phase, leading to today’s bacteria, and the eukaryotic phase, characterized by a much higher degree of organization and represented in the present world by a variety of microorganisms called protists. The eukaryotic cell spawned the Age of Multicellular Organisms, with its new principles of cellular association, differentiation, patterning, communication, and collaboration. To this age belong all plants, fungi, and animals, including humans, with each group itself organized hierarchically along an ascending scale of com¬ plexity, exemplified at each level by extant organisms. Finally, there is the Age of the Mind, with all its social and cultural implications and its attending moral responsibilities. In the following chapters, I shall take the reader through each of these succes¬ sive ages. I shall conclude with the Age of the Unknown, which encompasses the future of life and its timeless aspects.

Chapter 1

The Search for Origins

Virtu ally all the organic matter in the living world can be summarized sym¬ bolically, if not euphonically, by the formula CHNOPS, which stands for carbon (C), hydrogen (H), nitrogen (N), oxygen (O), phosphorus (P), and sulfur (S). These six elements, in myriad molecular combinations, make up the bulk of living matter. They were the main actors in the chemical birth of life as well. In order to reconstruct this momentous event, we must find out in what form the six biogenic elements were present on the primitive Earth and how, driven by the special physical-chemical circumstances that prevailed, they were first caught in a spiral of increasing complexity out of which life was born. First, what do we know of the setting in which life arose?

THE SETTING The Earth, four billion years ago, was beginning to recover from the battering by ce¬ lestial bodies that accompanied its violent birth.1 It had cooled sufficiently for water to condense on its surface. Islands were rising in the primeval oceans and starting to merge into continents. The lands were barren and the waters lifeless, but the scene was far from calm. Still in the throes of intense volcanic activity, the young Earth was pitted by red-hot craters spewing thick clouds of dust and fumes. It was ravined by deep cracks through which water seeped down to the molten core, later to erupt back up again, pressurized, super-heated, and laden with vapors extracted from the seething lava. Think of Yellowstone National Park, or of the solfataras of Sicily, the Hekla region in Iceland, the flanks of Mount Fuji in Japan, or the hot springs of Rotorua in New Zealand. One memory invariably comes to mind: the smell! The allpervasive stench of rotten eggs, the characteristic odor of hydrogen sulfide. Indeed, there is every likelihood that the cradle of life reeked of hydrogen sulfide. This fact has rarely been taken into account in origin-of-life scenarios. It deserves to be.

16

THE AGE OF CHEMISTRY

There was no oxygen in the atmosphere around the Earth four billion years ago. Free oxygen is a product of life. This is as close to certain as anything can be in sci¬ ence. In consequence, the state of many minerals was very different from what it is today. This is particularly true of iron. Leave an iron object outside for some time and it turns to rust, which results from the interaction of iron with wet oxygen. There was no rust, that is, no iron oxide, on the prebiotic Earth. Instead, iron was abundantly present in the oceans in a form, called ferrous, that is not found today because it would react immediately with atmospheric oxygen. The composition of the primitive atmosphere is still the object of debate. The prevalent view for a long time, popularized by the celebrated Urey-Miller experi¬ ment (see pp. 18-19), was that the atmosphere consisted of hydrogen (H2), methane (CH4), ammonia (NH3), and water vapor (H20), and thus was very rich in hydro¬ gen. This is now seriously doubted. According to many experts, carbon was proba¬ bly not present in combination with hydrogen (methane) but with oxygen (mostly carbon dioxide, or CO,). Nitrogen most likely existed as molecular nitrogen (N2) or in one or more of its associations with oxygen, not as ammonia. Traces, at most, of molecular hydrogen were present. If these new estimates are correct, the source of the hydrogen needed for the formation of the first biomolecules raises a serious problem (see chapter 3, “The Case of the Missing Hydrogen”). Another problem we encounter when looking at the prebiotic scene concerns phosphorus. This element, in the form of phosphate, is a conspicuous constituent of many important biomolecules, in particular nucleic acids. The fact is surprising, since phosphate is hard to find in the present world, at least in solution. The Earth contains plenty of phosphate, but locked up in water-insoluble calcium phosphate, the constituent of the mineral apatite. In marine and fresh waters, the concentration of phosphate is very low; indeed its availability is often a limiting factor in sustain¬ ing life in these environments. This was made clear when phosphates were added to cleaning powders. Contamination of lakes by phosphate-containing waste water led to eutrophication, an excessive proliferation of algae nurtured by the newly avail¬ able phosphate, which alters the food chain in such a way that oxygen becomes scarce and animal life severely hampered. How the rare phosphate molecule turned out to play its central biological role is an intriguing question. A possible answer to the problem could be acidity, a physi¬ cal property that is associated with sourness when mild—think of vinegar or lemon juice—and with bitingly corrosive powers when strong—think of aqua fortis (ni¬ tric acid), used in etching, or of vitriol (sulfuric acid), the favorite weapon of betrayed Victorian ladies. Apatite readily releases phosphate when exposed to even a mild acidic medium. Perhaps the primeval waters in which life originated had this property.2 What was the temperature of the prebiotic world? Little solid evidence is avail¬ able. This is unfortunate, because temperature is a critical parameter that severely limits the life span of relatively fragile biomolecules such as proteins, nucleic acids, and many of their building blocks. With this fact in mind, many chemists concerned

THE SEARCH FOR ORIGINS

17

with the origin of life have opted in favor of a cold environment, even below freez¬ ing point.3 Geochemists, on the other hand, do not favor a cold prebiotic world. Their esti¬ mates are in the upper r^nge, near the temperature of boiling water or even higher, compensated by a sufficiently high atmospheric pressure to keep the oceans from boiling. High temperature and high pressure are typical characteristics of underwater hydrothermal vents of the kind that have been discovered in several deep-lying areas of present-day oceans and that were no doubt more abundant on our young, volcani¬ cally convulsed planet.4 Also favoring a hot cradle for life is the finding that the most ancient organisms, according to comparative sequencing, are bacteria living in such vents or in volcanic springs at temperatures of up to 110°C (230°F). How about sunlight? The sun was cooler four billion years ago. It sent out about 25 percent less light energy to the Earth than today. However, this was probably offset by the greenhouse effect of atmospheric carbon dioxide, which may have been as much as one hundred times more abundant in prebiotic times than today. Ultraviolet radiation most likely was strong despite a cooler sun because, there being no oxygen, there was no ozone shield (ozone is made of three oxygen atoms). One more point deserves to be mentioned about the prebiotic environment: It was devoid of life. This sounds like a tautology but it holds an important implica¬ tion, already noted by Charles Darwin more than a century ago. In an oft-quoted letter to a friend, Darwin wrote: “It is often said that all the conditions for the pro¬ duction of a living organism are now present, which could ever have been present. But if (and oh what a big if) we could conceive in some warm little pond, with all sorts of ammonia and phosphoric salts, light, heat, electricity, etc., present, that a protein compound was chemically formed ready to undergo still more complex changes, at the present day such matter would be instantly devoured or absorbed, which would not have been the case before living creatures were formed.”5 As a rule, this passage is cited for its reference to a “warm little pond.” But Darwin makes a very pertinent statement. There was nothing in the prebiotic world that might “biodegrade” organic molecules. These could survive and accumulate for a very long time, subject only to much slower physical and chemical degradation. Summing up, we may state first that there was plenty of water on the Earth when life originated. It could hardly have been otherwise, since water is the vital element par excellence. Wake up in the desert after a night sprinkle and you witness the mir¬ acle of water. Everywhere, seeds left to dry in the barren soil spring to life in an enchantment of colors. This does not mean that the prebiotic waters would have been our preferred spot for a South Sea vacation. They were probably scaldingly hot, perhaps bitingly acidic, and loaded with ferrous iron, phosphate, and other minerals torn from the inner depths of the Earth. The atmosphere over the waters was heavy with carbon dioxide, nitrogen, hydrogen sulfide, and water vapor, but most likely poor in hydrogen. A pale but unshielded sun filtered through to bathe the surface waters in ultraviolet radiation, light, and warming infrared rays trapped by the carbon dioxide cover.

THE AGE OF CHEMISTRY

18

On such a planet, two possible settings existed for life to unfold: shallow surface waters where the “soup” could thicken and “cook” under sunlight; or dark, deepseated hydrothermal vents harboring strange chemistries. Or, perhaps, currents between the two settings allowed some steps of biogenesis to occur in one environ¬ ment, and others in the other.

CHEMISTS ON THE TRAIL Once information started to become available concerning the physical-chemical setting in which life arose, the next obvious step was to re-create conditions in the laboratory under which the first stages in the origin of life could be reproduced. The Soviet biochemist Alexander Oparin is generally credited with originating this new field of inquiry. He first published a booklet on the origin of life in 1924 and later expanded it to a full-size book that went through a number of revised editions, some of which were translated into English.6 Largely inspired by the cellular theory of life and by what was known in his days as colloid chemistry, Oparin’s concept of the origin of life looks naive today. But he had the great merit of actually testing his ideas in the laboratory, where he prepared and studied a number of molecular aggregates that he saw as possible precursors of the first cells. For a long time, Oparin did not attract much of a following. The feeling was—it certainly was my opinion at the time—that it did not make much sense searching for the origin of something that was so poorly understood. Things changed in the early 1950s. On April 23, 1953, the British magazine Nature published a brief note entitled “A Structure for Deoxyribose Nucleic Acid,” by the American James D. Watson and the Englishman Francis Crick.7 This epoch-making paper, which earned its authors a Nobel Prize nine years later, introduced the now famous double helix, which has become a symbol of the recent revolutionary advances in our understanding of life. Three weeks later, an equally brief and momentous note, titled “A Production of Amino Acids under Possible Primitive Earth Conditions, appeared in the May 15, 1953, issue of the magazine Science, the American coun¬ terpart of Nature. Written by a young graduate student, Stanley L. Miller, this papei inaugurated modern research on the origin of life.8 Miller worked in the Chicago laboratory of Harold Urey, a physicist who was awarded the 1934 Nobel Prize in chemistry for the discovery of heavy hydrogen, or deuterium. In later years, Urey became interested in the formation of the planets.9 It was he who defended the view that the early Earth’s atmosphere was a hydrogenrich mixture of molecular hydrogen, methane, ammonia, and water vapor. Miller decided to find out how lightning might have affected such an atmo¬ sphere. With the reluctant consent of his mentor, who considered this project too iffy for a doctoral thesis, he simulated primitive thunderstorms by producing repeated electric discharges inside a sealed glass enclosure containing a gaseous

THE SEARCH FOR ORIGINS

19

mixture of methane, ammonia, and hydrogen, through which water was continually recycled by evaporation and condensation, as would have happened over a primeval ocean. Imagine young Miller’s surprise when he saw the “ocean” assume a pinkish glow in a matter of a few days; his eagerness when he opened the con¬ tainer and removed the water for chemical analysis; his delight when the analysis showed the presence of several amino acids and other organic molecules typically found in living organisms. The results exceeded his wildest expectations and pro¬ pelled him instantly into the firmament of celebrities. This historic experiment alerted organic chemists to the origin of life as a chemi¬ cal problem. Itjsparked the birth of a new discipline, termed abiotic (without life) or prebiotic (before life) chemistry, concerned with the spontaneous formation of bio¬ logical substances under conditions that might have prevailed on our planet some four billion years ago. Many important molecules have indeed been obtained in this fashion, though frequently under conditions somewhat more contrived than one would like for a truly abiotic process.10 In this rich crop, Miller’s original experi¬ ment remains a paradigm, virtually the only one conceived exclusively with the aim of reproducing plausible prebiotic conditions, with no particular end product in mind. Ironically, the relevance of the conditions of this experiment is now seriously questioned. The prebiotic atmosphere may well have been much less rich in hydro¬ gen than Urey thought. As Miller himself found out, if, in agreement with current views, methane is replaced by carbon dioxide, and ammonia by molecular nitrogen, in the gas mixture he used in his celebrated experiments, and if molecular hydrogen is left out, the yield of organic substances falls practically to zero. The verdict is not in yet. Estimates of the composition of the early atmosphere are very uncertain and may be revised again in the future. In the meantime, unexpected support for the validity of Miller’s findings, if not his experimental conditions, has come from outer space.

SEARCHING THE SKIES One of the most powerful techniques used to probe the cosmos is spectroscopy. Put simply, this technique allows analysis of incoming light after it has been separated into its component wavelengths by passage through a prism—in the same way that sunlight is broken into a rainbow of colors (wavelengths) by droplets of water. With appropriate decomposers and amplifiers, the same technique can be extended to nonvisible kinds of electromagnetic waves, such as ultraviolet, infrared, or radio waves, even of very low intensity. Substances present in outer space act as filters that absorb radiations of certain characteristic wavelengths (colors or their equiva¬ lent). In consequence, the absorbed radiations are found to be missing or attenuated in the recorded spectra (dark bands in the rainbow). Alternatively, certain wave-

20

THE AGE OF CHEMISTRY

lengths may be enhanced by emission from energetically excited substances. In many cases, the substances causing the absorption or emission can be identified from the spectral patterns, which serve as fingerprints of the substances involved. The analysis of microwave radiation—the kind that, at much higher intensities, heats ovens—has turned out to be particularly fruitful in this respect. Spectroscopic probing has revealed that the cosmic spaces are permeated by an extremely tenuous cloud of microscopic particles (interstellar dust) containing a number of potentially biogenic molecules, mostly highly reactive combinations of carbon, hydrogen, nitrogen, oxygen, and, sometimes, sulfur or silicon that would hardly remain intact under Earth conditions but could give rise to biologically sig¬ nificant compounds.11 This presumably happens in the formation of comets. Long seen as fiery objects hurtling through space while scattering a stream of sparks behind them, comets really consist mostly of dust and ice loaded with a variety of organic compounds. This has been learned by spectral analysis and, thanks to the recent passage in the Earth s neighborhood of the famous comet discovered in 1681 by the English astronomer Edmund Halley, by direct chemical tests with the help of instruments carried by a spacecraft. Even more solid is the evidence brought to us by meteorites. For example, the Murchison meteorite, which fell in 1969 in Murchison, Australia, was found to con¬ tain a number of amino acids remarkably similar in nature and in relative quantity to those obtained by Miller in his experiments. This kind of evidence, while provid¬ ing a considerable boost to the significance of Miller s results, tells us further that organic compounds can survive the scorching crash of celestial bodies through the atmosphere. There is thus ample evidence that a number of biogenic compounds can form spontaneously under primitive Earth conditions, in interstellar space, and on comets and meteorites. Most likely, such compounds provided the first seeds of life. How much was made locally, how much was brought in from outer space, is still widely debated. The Belgian-born American astrophysicist Armand Delsemme, from the University of Toledo, believes that virtually all the building blocks of life, as well as all the terrestrial water, were carried to the Earth by comets that contributed to the final accretion of our planet.12 According to Miller, on the other hand, the chem¬ ical precursors of life were formed mostly on Earth itself.

THE FIRST STEPS Given the evidence from simulation experiments on Earth and the analysis of extraterrestrial objects, complemented by plausible conjectures, the following sce¬ nario may reasonably be proposed for the birth of life on Earth some four billion years ago. The seeds of life arose in space and in the atmosphere, in the form of various combinations of carbon, nitrogen, hydrogen, oxygen, and, as we shall see

THE SEARCH FOR ORIGINS

21

later, sulfur. Under the influence of electric discharges, radiation, and other sources of energy, the atoms in these combinations were reshuffled to produce amino acids and other basic biological building blocks. Brought down by rainfall and by comets and meteorites, the products of these chemical reshufflings progressively formed an organic blanket around the lifeless surface of our newly condensed planet. Everything became coated with a carbonrich film, openly exposed to the impacts of falling celestial bodies, the shocks of earthquakes, the fumes and fires of volcanic eruptions, the vagaries of climate, and daily baths of strong ultraviolet radiation. Rivers and streams carried these materi¬ als down to the seas, where the materials accumulated until “the primitive oceans reached the consistency of hot dilute soup,” to quote a famous line from the British geneticist J. B. S. Haldane.13 In rapidly evaporating inland lakes and lagoons, the soup thickened to a rich puree. In some areas, it seeped into the inner depths of the Earth, violently gushing back as steamy geysers and boiling underwater jets. All these exposures and churnings induced many chemical modifications and interac¬ tions among the original components showered from the skies. The major outcome of all this geological cookery most likely consisted of some sticky, brownish, water-insoluble goo of indefinite composition, only too familiar to organic chemists who invariably see it staining the walls of their flasks when something goes wrong in their concoctions. It coated Miller’s vessel as well but hardly seemed to him worth mentioning since life could not possibly arise out of such material. Somewhere on the primitive Earth, however, the seeds of life were saved from turning into goo and were channeled in the direction of productive chemical complexification. What was this direction? The most widely accepted answer to this question is not what one would expect. Consider the following three statements, which all happen to be true: (1) Amino acids are among the most conspicuous products of abiotic chemistry, both on Earth and in space. (2) Amino acids are the building blocks of proteins. (3) Proteins are centrally important biological constituents. Most important for our purposes, the vast majority of enzymes, the catalytic agents responsible for the chemical reac¬ tions that take place in living organisms, are made of proteins. What conclusion do you draw? I can hear the whole class answering in triumphant unison: The next step on the way to life was the formation of proteins, which, in turn, provided the first enzymes whereby the biogenic process further unfolded. Right? Wrong! At least according to current majority opinion. Proteins, it is argued, must have been preceded by ribonucleic acid (RNA). The main reason for this belief is that, in the present living world, RNA molecules provide both the catalytic machinery and the information—itself derived from deoxyribonucleic acid (DNA), as we shall see later—for the assembly of amino acids into proteins. No doubt, those in the class with a little biochemical savvy will immediately note the flaw in this argument. In the present living world, no RNA molecule could arise without the help of protein enzymes. In other words, protein makes RNA, which makes protein, which makes RNA . . . , and so on. What came first: protein or RNA? It is

THE AGE OF CHEMISTRY

22

the old “chicken or egg” problem, which, it is said, kept a Chinese mandarin pon¬ dering fruitlessly all his life. Molecular biologists have escaped this sorry fate by calling on Crick’s “Central Dogma”—not a real dogma, of course, but a logically derived postulate that the co¬ discoverer of the double helix enunciated as early as 1957, when little of the empiiical evidence that now overwhelmingly supports it was available. This postulate states that information tlows only from nucleic acids to proteins, never in the reverse direction.14 Hence, RNA came before protein. This affirmation gained a great boost in the early 1980s when two American investigators, Thomas Cech from the University of Colorado at Boulder and Sidney Altman from Yale Univer¬ sity, who shared a Nobel Prize in 1991 for their discovery, found independently that certain RNA molecules were endowed with catalytic activity.15 This fact suggested that RNA enzymes—called ribozymes by Cech—could have done the catalytic work of proteins in what is now known as the “RNA world,” an expression coined in 1986 by the Harvard chemist Walter Gilbert,16 whose method of DNA sequenc¬ ing gained him a Nobel Prize in 1980. Gilbert defines the RNA world as a stage in the early development of life in which “RNA molecules and cofactors [were] a suf¬ ficient set of enzymes to carry out all the chemical reactions necessary for the first cellular structures.”17 I shall have more to say on this topic. In the meantime, we may take it as most likely that, whatever came before RNA, the molecules ancestral to present-day pro¬ teins came after. The evidence supporting a primary role of RNA in the birth of these molecules leaves little room for doubt, as will become clear in part II. If we accept this premise, we must now face the chemical problems raised by the abiotic synthesis of an RNA molecule. These problems are far from trivial.

THE WAY TO RNA RNA molecules are long, chainlike assemblages made of a large number—up to many thousands—of units called nucleotides. Each nucleotide consists of three parts: phosphate, ribose, which is a 5-carbon sugar, and a base, of which there are four different kinds—adenine, guanine, cytosine, and uracil. All four are flat, ringshaped molecules of some complexity, made of carbon, nitrogen, hydrogen, and (except for adenine) oxygen atoms. Adenine and guanine belong to the group of purines, built from two fused molecular rings. Cytosine and uracil are pyrimidines, with a simpler, one-ringed structure. In RNA, the nucleotides are linked by bonds between the ribose of one and the phosphate of another. As a result, all RNA molecules share the same backbone (except for its length) of alternating phosphate and ribose molecules. The bases are attached to each ribose unit of this backbone, as shown in the following diagram, in which each boxed unit is a nucleotide:

THE SEARCH FOR ORIGINS

23

— phosphate-ribose - phosphate-ribose

0

ba se

phosphate-ribose

base

base

Chemists have achieved some success in making each of the five organic com¬ ponents of RNA, but with poor yields and under conditions far removed from a plausible prebiotic setting and different for each substance. Hitching the compo¬ nents together in the right manner raises additional problems of such magnitude that no one has yet attempted to do so in a prebiotic context. In any RNA molecule, the bases provide the informational part. They are the four letters with which RNA words are constructed. The phosphate-ribose backbone, on the other hand, plays a purely structural role. Accordingly, one line of research has been to look for simpler backbones to carry the same bases and to support the same kind of information. Orgel and his coworkers at the Salk Institute have pursued this line particularly diligently and ingeniously. They have produced a number of inter¬ esting molecules but have not yet solved the problem to their own satisfaction.18 Re¬ cently, a young Danish investigator, Peter Nielsen, has attracted considerable interest with a molecule he calls peptide nucleic acid (PNA), in which the backbone consists of amino-acid derivatives strung together as in proteins.19 No evidence is yet avail¬ able to evaluate this intriguing way of combining the chicken with the egg. None of these molecules, if they ever existed, have left traces in extant organ¬ isms. Furthermore, it is far from clear how they could have arisen and how they could subsequently have been replaced by authentic RNA molecules. It is fair to state that no mechanism has yet been found that could account satisfactorily for the prebiotic synthesis of RNA, despite considerable efforts by some of the best chemists in the world. Even the staunchest defenders of the RNA world have expressed despondent views on the future prospects of this line of research.20 Could chance be the answer? All thirteen spades in the same hand? A single, highly improbable combination of circumstances that led to the spontaneous forma¬ tion of a few RNA molecules somewhere in the prebiotic world? This possibility has been evoked on the grounds that self-replication could then have ensured the propagation of the RNA. The RNA world would thus have been born from a single molecular seed, itself the product of a chance event. This explanation does not hold water. In replication, existing RNA provides only information. The actual making of new RNA molecules requires the same chemical complexity as that of the first. We need thirteen spades many times in succession. This is all the more true because the RNA world was not a fleeting, transient moment in the history of life. It was a long, drawn-out period that lasted all the time that was needed for a protein-synthesizing machinery to arise and turn out the vari¬ ous protein enzymes that eventually took over the job of sustaining emerging life catalytically. We don’t know whether this took millennia or years, but it certainly took a long enough time to be possible only with a sturdy chemical underpinning.

24

THE AGE OF CHEMISTRY

THE LESSON OF METABOLISM The conclusion is clear. We need a pathway, a succession of chemical steps leading from the first building blocks of life to the RNA world. Chemistry, however, has so far failed to elucidate this pathway. At first sight, the kind of chemistry needed seems so unlikely to take place spontaneously that one might be tempted to invoke, as many have done and some still do, the intervention of some supranatural agency. Scientists, however, are condemned by their calling to look for natural explanations of even the most unnatural-looking events. They must even, in the present case, eschew the facile recourse to chance, as I hope to have made clear. The pathway to life must have been downhill all the way, with at most a few rare humps that could be negotiated with the help of the acquired momentum. One would expect such a roadway to be readily visible. Yet, so far, like some artfully hidden jungle trail, it has eluded every search, despite extensive experimentation and much imaginative theorization and speculation. Many devotees of abiotic chemistry, encouraged by the numerous positive results that have already been achieved, continue to believe that further attempts at reproducing early synthetic reactions in the laboratory will eventually uncover the trail. Others, howevei, impressed with the complexity of the many molecular assemblages needed for the continuing operation of even a skeletal RNA world, tend to be less sanguine. A pathway does exist, of course, for everyone to see. It is being followed in every corner of the Earth by trillions of trillions of living cells. The green cells of plants and trees and many bacteria do so even without the celestial manna that seeded life in the beginning. These cells construct all their constituents from such simple materials as carbon dioxide, water, nitrate, sulfate, and traces of a few other mineral salts. These pathways make up metabolism. They are known in great detail. Why look elsewhere when nature points the way? No reason, except that nature’s way appears so strange and tortuous to the chem¬ ically trained mind that one cannot avoid the feeling that a simpler and more straightforward way must have existed before. Impressions can be deceptive, how¬ ever. If life did start along pathways that have nothing to do with those of presentday metabolism, why were the early pathways replaced by new ones? And, espe¬ cially, how? Biologists have a ready answer to the first question. They simply assume that the new pathways were better than the old ones, and they call on nat¬ ural selection, the universal motor of biological evolution and progress, to promote the change. The second question, however, is not as easily answered. It simply won't do to visualize a brand-new network of pathways as developing independently of the old ones and taking over only after being completed. That is what we do when we superimpose a new railroad or superhighway network over a primitive road system. But we do it with the kind of prescience and design that we have agreed the bio¬ genic process did not enjoy. The naturally occurring replacement of prebiotic path-

THE SEARCH FOR ORIGINS

25

ways by present-day pathways must necessarily have taken place gradually, one step at a time. It demands some sort of congruence between the early and the later pathways.21 To understand this point, imagine an old road map dating back to horse and buggy days. It consists of a network of roads linking the various cities and villages of the country. Suppose now that a benevolent contractor builds a better road lead¬ ing, say, from site A to site B. Clearly, his gesture will be wasted if sites A and B are not parts of the old network. What would be the use of a road linking two points that each are in the middle of nowhere? On a (proto)metabolic map, the cities and villages are chemical intermediates, and the roads between them, drawn as arrows, represent chemical transformations of one or more intermediates into others. Most often, the connecting arrows indicate the presence of a catalyst, or enzyme, respon¬ sible for the transformation. On such a map, the new road built by a benevolent contractor would correspond to a new enzyme—catalyzing, say, the conversion of intermediate A to intermediate B—arising by chance through the operation of the protein-synthesizing machinery of the RNA world. This enzyme, part of the new metabolic network that is replacing the old protometabolic network, will be useless, and, therefore, will not be retained by natural selection, if it does not fit within the old network, that is, if intermediates A and B are not parts of the protometabolic map. This is the essence of my argument in favor of congruence. It will become clearer after we have considered the mechanism of selection in part II.

Chapter 2

The First Catalysts of Life

From the first building blocks of life well into the RNA world. Such is the

hidden trail we must try to uncover. We know the outcome: complex organic sub¬ stances made of carbon, nitrogen, oxygen, hydrogen, and phosphorus atoms linked with each other in molecular structures that are known with great accuracy. Our task is to find out how such arrangements arose naturally from simpler arrange¬ ments of the same atoms present in the prebiotic environment. Our major clue comes from the requirement for congruence. The metabolic maps reproduced in all biochemistry textbooks are modernized versions of the ancient networks and should help us in our task. Enzymes are the signposts on metabolic maps. Virtually every one of the thou¬ sands of chemical reactions that take place in any living cell is catalyzed by an enzyme. Most of these reactions would not occur at all without enzymes. Hundreds of fatal or severely disabling genetic diseases characterized by a single enzyme deficiency attest to this fact. It is very unlikely that protometabolism could have done without catalysts. If these were not proteins, what were they?

CATALYSIS WITHOUT PROTEINS Just as metabolism cannot operate without enzymes, protometabolism, whatever pathways it may have followed, could not have functioned without catalysts. Text¬ books define catalysts as substances that specifically help reacting molecules to get together and interact, but are not themselves consumed in the reaction and, there¬ fore, can serve an indefinite number of times in succession. Why did emerging life need catalysts? Chemical reactions take place all the time in the physical world without the help of catalysts. There are two reasons for the requirement for catalysts: rapidity and yield. Uncatalyzed reactions most often are very slow. This means, in the prebiotic set-

THE FIRST CATALYSTS OF LIFE

27

ting, that important reaction products might well have been destroyed almost as rapidly as they were made, never reaching a sufficient level to participate in a sub¬ sequent step. Without catalysts to speed up reactions, protometabolism would have been in the sorry predicament of the fifty Danaid sisters condemned to pour water eternally into a bottomless barrel. Low yield raises an equally thorny problem. Take the following case. In 1960, the Catalonian-born American chemist Juan Oro discovered that adenine, one of the constituents of RNA, could form in a single step from ammonium cyanide, which is considered a compound likely to have been present on the prebiotic Earth.1 This remarkable finding of a central biomolecule arising in such simple fashion was hailed as almost equivalent to Miller’s historic experiment and has become a text¬ book paradigm of the power of abiotic chemistry. Yet the highest yield of adenine in this reaction was 0.5 percent, which means that 99.5 percent of the reaction mix¬ ture consisted of other stuff. Another example is the synthesis of ribose, another RNA component, from formaldehyde—also a classic of abiotic chemistry. In this case, the yield is less than 0.1 percent, and at least forty other sugars are present in the mixture.2 Such low yield of any one product is typical of many organic-chemical reactions when they are not subjected to the kind of strict constraints imposed by chemists in the laboratory. There are always side reactions, and these are all the more numerous the looser the constraints. Problems of this sort are compounded when a process requires several successive steps. Imagine a short sequence of only three steps— from A to B, from B to C, and from C to D—each with a yield (high for a prebiotic reaction) of 1 percent. In terms of A, the yield of B will be 0.01, that of C, 0.0001, and that of D, 0.000001, or one in one million. Even under the best conditions, chemists have to fight against this kind of vanishing act. They often purify a signif¬ icant intermediate between two steps and change conditions at each step to maxi¬ mize yield. Prebiotic chemists are acutely aware of these difficulties and have come up with a number of more or less plausible mechanisms whereby relevant interme¬ diates could have been selectively concentrated from the prebiotic soup. Let us apply the rule of congruence and ask how nature copes with the problem. Nature’s solution lies in enzyme specificity. Biological catalysts are truly remarkable in this respect. They are by far superior to the best catalysts devised by human ingenuity; as a consequence, the large-scale production and engineering of enzymes for industrial purposes has become an important branch of modern biotechnology. Many enzymes catalyze a single reaction or a set of closely similar reactions that would hardly occur without a catalyst. Enzymes are proteins or, exceptionally, RNA molecules (ribozymes); they could not have been present in a pre-RNA world that had not yet developed the RNAbased machinery of protein synthesis. We must look elsewhere for the catalysts of protometabolism. To most investigators, elsewhere has meant the inorganic mineral world, since we are talking of a time when the organic world was still in its infancy. In contemporary biochemistry, many enzymes act with the help of an auxiliary sub-

28

THE AGE OF CHEMISTRY

stance of inorganic nature, most often an atom of a metal such as iron, copper, cal¬ cium, magnesium, zinc, molybdenum, cobalt, or manganese. A difficulty with metals is that they often need some supporting structure—usu¬ ally a protein framework—to interact productively with the molecules that partici¬ pate in the reaction they help catalyze. Thus, a great deal of attention has also been given to mineral surfaces likely to provide the necessary frameworks and also, per¬ haps, to act catalytically by themselves. Favorites are clay particles, already advo¬ cated more than fifty years ago by the British physical chemist John Desmond Bernal, one of the pioneers in the study of the origin of life.3 Clays come in many different microcrystalline forms, and some, indeed, display catalytic activity. Montmorillonite, for example, which owes its name to the French town of Montmorillon, near which it is quarried, has been found to facilitate the assembly of short RNAlike chains from suitably prepared nucleotides.4 Unlike metals, however, clays, which are aluminum silicates, have left no trace in present-day life to suggest that they played any role as catalysts in protometabolism. As an alternative to clays, Gustaf Arrhenius, from the Scripps Institution of Oceanography in La Jolla, California, has advocated a catalytic role for what he calls “positively charged, double layer hydroxide minerals,” especially in the syn¬ thesis of sugar phosphate molecules.5 The German chemist and patent attorney Gunter Wachtershauser has con¬ structed an elaborate model—by far the most detailed such model on record—of the development of protometabolism on the surface of pyrite crystals.6 Known as fool’s gold because of its golden tinge, pyrite is a mineral composed of iron and sulfur. In Wachtershauser’s model, the mineral owes its catalytic role to the fact that electrically charged objects attract each other if they bear charges of opposite sign (and repel each other if they bear charges of the same sign). Pyrite, which is positively charged, offers a surface on which, according to the German author, neg¬ atively charged molecules bound by electrostatic attractions could be brought into close neighborhood with each other and caused to interact in various ways. Inciden¬ tally, Wachtershauser accounts for the importance of phosphate in metabolism by his model: Phosphate is negatively charged and thus allows the molecules to which it is attached to bind to the pyrite surface. Chemically, Wachtershauser’s model is largely inspired by present-day metabo¬ lism and obeys the congruence rule. But some of the mechanisms invoked are spec¬ ulative and need experimental testing. Furthermore, his catalyst lacks specificity, except for the very broad property of binding, with variable strength, anything that is negatively charged. The model relies heavily on autocatalysis. A number of other workers have appealed to this concept as a solution to the problem of prebiotic catalysis.7 Autocatalysis occurs when the product of a chemical reaction is catalyti¬ cally helpful to the reaction: B catalyzes the conversion of A to B. A slow-starting reaction may thereby progressively accelerate, sometimes to the point of becoming explosive. There is no question of emerging life ever exploding, but the idea is that processes difficult to initiate can become self-sustaining by autocatalysis once they

THE FIRST CATALYSTS OF LIFE

29

are primed. This is a way of catching chance events and turning them into “going concerns.” All the mechanisms mentioned so far may well have played a role in proto¬ metabolism. But could they have accomplished the whole job without the help of proteins? Several workers have expressed serious doubts about this and insisted on the early participation of protein catalysts, at the risk of contravening the Central Dogma.8 There is much to say for this view, especially if the term “protein” is replaced by “peptide,” defined as any chainlike assemblage of amino acids, not just the particular kind made from twenty specified varieties of amino acids by an RNA-dependent machinery.

THE CASE FOR PREBIOTIC PEPTIDES Amino acids, the building blocks for making peptides, were probably among the earliest organic substances to be present in the prebiotic world. More than twelve kinds of amino acids formed in significant amounts in Miller’s flask, and the same kinds have been extracted from meteorites. Some of these amino acids are found in proteins today, others are not. No matter, all possessed the essential characteristics that allow amino acids to join into peptides: the carboxyl group (-COOH) responsi¬ ble for the acid nature of the substances, and the ammonia-derived amino group (-NH2). In peptides and proteins, these two groups are joined to form a peptide bond (-CO-NH-), with the release of one water molecule. Could the primeval amino acids have joined into peptides under prebiotic condi¬ tions? What looked like a simple positive answer to this question was found in 1958 by the American biochemist Sidney Fox, long of the University of Florida, now at the University of South Alabama.9 His recipe: Just heat a dry mixture of amino acids for three hours at 170°C (338°F). Water comes out and you get a plas¬ tic-like solid that, when ground and mixed with water, yields up to 15 percent of its weight as a water-soluble product made, on average, of some fifty amino acids joined together. To this product Fox gave the name proteinoid, a cautious choice since proteinoids are far from having the regular chainlike structure of peptides. For Fox, this discovery initiated a lifelong avocation. He found that proteinoids spontaneously form microscopic vesicles, or “microspheres,” which he saw as the first cells, and spent his whole career pursuing these studies. Few origin-of-life experts are as sanguine as Fox concerning the significance of his results. It has been objected that the conditions required for the formation of proteinoids are not likely to have obtained on the prebiotic Earth, that the resulting material has more in com¬ mon with primeval “goo” than with proteins, and that the microspheres are a far cry from anything that could be called a cell. I tend to share these misgivings but retain as possibly significant two of Fox’s findings: proteinoids possess some weak,

30

THE AGE OF CHEMISTRY

enzyme-like catalytic properties; and the amino-acid composition of proteinoids is specific and reproducible despite the disordered conditions of their formation. This means that the bonds between amino acids did not form purely at random, but that certain associations were advantaged and others excluded. A more orthodox way of getting peptides was discovered in 1951, before Fox’s findings, by a German chemist, Theodor Wieland.10 At that time, biochemists had just discovered the thioester bond, which turned out to be of cardinal importance in all present-day living organisms, probably also in the origin of life. This excep¬ tional situation warrants a brief excursion into biochemistry.

INTRODUCING THIOESTERS An ester arises from the joining of a hydroxyl group (-OH), characteristic of alco¬ hols, with a carboxyl group (-COOH), characteristic of organic acids. One mole¬ cule of water is removed in the process, and the two building blocks are linked by what is called an ester bond (-O-CO-). A thioester arises similarly from the joining, with the removal of water, of a thiol with an acid. Thiols (from the Greek theion, sulfur) are the equivalent of alcohols in which the oxygen atom is replaced by one of sulfur. They are characterized by the thiol group -SH. Thioester bonds have the structure -S-CO-. Wieland became interested in thioesters as a student of Feodor Lynen, the dis¬ coverer of the first known natural thioester, a compound of acetic acid with a thiol designated coenzyme A in biochemical jargon.11 Coenzyme A, a molecule of cen¬ tral importance, was discovered in 1947 by the German-born American biochemist Fritz Lipmann, the “father” of bioenergetics, who received the 1953 Nobel Prize in medicine for this discovery. Lynen, who was similarly rewarded in 1964, found that thioesters are the natural intermediates in the synthesis of esters from acids and alcohols. The main problem in the making of an ester from an alcohol and an acid is that a molecule of water needs to be extracted. Such a reaction—the closure of a bond with the release of water—is called a condensation. Condensation reactions do not take place spontaneously in an aqueous medium because there is too much water around. The spontaneous direction of the reaction—that is, the direction that does not cost any energy but, on the contrary, releases energy—is the reverse of the condensation, the splitting of the bond with the help of water, or hydrolysis. For example, in the presence of a suitable enzyme, esters are hydrolyzed into alcohols and acids. For the reverse process of condensation of alcohols and acids into esters to occur, energy must be spent, the water molecule must be forcibly extracted. Chemists do this with special reagents called condensing agents. Nature uses a different device. It starts— spending energy in the process—by condensing the acid with a thiol (coenzyme A) into a thioester. This is the energy-requiring, water-extracting step. In a second step.

THE FIRST CATALYSTS OF LIFE

31

the acid is transferred from coenzyme A to the alcohol, and coenzyme A is released, ready to participate in a new round. Group-transfer reactions of this kind play a primordial role in the innumerable condensation reactions that underlie the biosyn¬ thesis of all complex biological molecules, including not only proteins and nucleic acids, but also carbohydrates, fats, and many others.12 Back to Wieland. As a firsthand witness of the discovery of the biological process of ester formation by group transfer from a thioester, he decided to find out whether this would also work for peptides, which likewise are formed by condensa¬ tion reactions, but between amino acids. So, Wieland synthesized amino-acid thioesters and simply threw them together in water. Amazingly, it worked! Peptides were formed, even though no catalyst was present.13 There is an amusing historical twist to this finding. When the mechanisms of protein synthesis were unraveled in the late 1950s and early 1960s, Wieland’s results were found to be irrelevant. Proteins are indeed formed by group transfer— that much remains true—though not from thioesters, but from esters (of amino acids and RNA molecules). Wieland’s vindication came a few years later, when Lipmann made the startling discovery that certain bacterial peptides—for example, the antibiotic gramicidin S—are synthesized in nature from thioesters.14 The thiol involved in this process was found to be pantetheine, which is itself the business end of coenzyme A, the central thiol discovered by Lipmann twenty years before. So do the mysterious wheels of science turn. In discussing his finding, Lipmann suggested that the thioester-dependent mech¬ anism of peptide formation may have preceded the RNA-dependent mechanism of protein synthesis in the development of life. I have adopted this suggestion and transposed it to the earliest steps of the biogenic process. For reasons that I shall explain in greater detail later, I believe that thioesters played a key role in the devel¬ opment of life. This belief fits with two central requirements of the trail we are try¬ ing to uncover: (1) congruence—thioesters are immensely important in present-day metabolism; and (2) the physical-chemical setting of the cradle of life—the thiol group is derived from hydrogen sulfide (H9S), the putrid but vital gas that pervaded the prebiotic world. It is my suggestion that thiols were part of the early organic molecules that seeded the development of life on the prebiotic Earth. Given the primeval setting, this sug¬ gestion appears eminently plausible, but the means of testing it have long been lack¬ ing because abiotic chemists, for reasons of their own, tended to shy away from sulfur chemistry. The omission has been repaired. A recent contribution from Miller’s laboratory describes a plausible procedure for the prebiotic synthesis of two natural thiols.15 One is coenzyme M, a metabolic cofactor of particularly ancient, methane-producing bacteria known as methanogens. The other is cysteamine, a con¬ stituent of pantetheine, which we have seen is the key component of coenzyme A and the natural cofactor involved in the synthesis of bacterial peptides. The Miller group has, in fact, been able to obtain the entire pantetheine molecule under plausible pre¬ biotic conditions.

THE AGE OF CHEMISTRY

32

I have made a further, more controversial suggestion, namely, that conditions on the prebiotic Earth favored the formation of thioesters from the primeval thiols and the amino acids and other acids that were presumably present also in large amounts. This possibility is more questionable because it involves the spontaneous occur¬ rence of an energy-requiring condensation reaction. I shall discuss this in the next chapter when considering the sources of protometabolic energy. Let us take it as a working hypothesis for the time being.

CATALYTIC MULTIMERS TO THE RESCUE Granting the presence of thioesters of amino acids, we know from Wieland’s results that peptides are likely to assemble spontaneously from these materials, even with¬ out a catalyst. In addition to peptide bonds, such assemblages could have included ester bonds as well, since hydroxy acids (with an alcohol group) were probably present also in large amounts in the primeval soup, according to Miller s results. Rather than peptides, therefore, which are made exclusively from amino acids, I have chosen to call the resulting molecules multimers.16 Why this linguistic monstrosity—which combines the Latin multus, many, with the Greek meros, part—and not the more orthodox polymer (from the Greek polys, many) or oligomer (from the Greek oligos, few)? Because polymet sounds too long, at least to me, and oligomer too short, and because both terms evoke images of regularity and homogeneity that I wish to avoid. The multimers of my model are a mixed bunch, containing more than a few building blocks, but fewer than the average polymer. My final suggestion, which many may find the most controversial, is that cata¬ lysts performing, albeit in crude form, the main activities carried out by enzymes in present-day metabolism were present in the multimer mixture and served as the main catalysts, or protoenzymes, of protometabolism. I have no proof of this, only some presumptions. According to my hypothesis, the multimers arose from random interactions among whatever thioesters were present. This does not mean that the resulting mix¬ ture was random, in the sense of containing all sorts of associations in a completely disordered fashion, without rule or reproducibility. On the contrary, we may take it that, as long as conditions remained the same, the mixture would have had a con¬ stant and reproducible composition, corresponding to a tiny subset of all the possi¬ ble associations that could be made from the available building blocks. A great many such associations would have been excluded either at the level of forma¬ tion—they were made too slowly or not at all—or at the level of breakdown—they were destroyed too quickly. Solubility in water would have been an additional selective factor, although it is conceivable that some molecules were catalytically

THE FIRST CATALYSTS OF LIFE

33

active in insoluble form. Finally, there is the possibility that catalytically active molecules were protected against breakdown by the molecules on which they acted, as many enzymes are by their substrates today. Only molecules that passed this multiple screening would'have made a significant contribution to the resulting mix¬ ture. Because of the strictly physical-chemical nature of the factors involved in the screening, the composition of the mixture would have stayed the same as long as the conditions did not change. This point is important. It makes this part of the bio¬ genic process reproducible and deterministic despite being dependent on random interactions. Whether this reproducible mixture would have included the protoenzymes required for protometabolism is conjectural but not implausible. The reasons for this are the following. First, we know that some of Fox’s proteinoids, in fact even single amino acids or mixtures of amino acids,17 can display crude catalytic activi¬ ties. The same should be true of the multimers I postulate. Next, the molecular con¬ figurations that would be expected to confer stability, such as a large enough mole¬ cular size and a compact or cyclic conformation, are also the configurations that a protein chemist would consider most likely to be required for catalytic activity. Third, as I shall explain in chapter 7, present-day enzymes must have started as rel¬ atively short peptides, probably no more than twenty to thirty amino acids long, perhaps much shorter. This fact makes the presence of catalytic molecules in the multimer mixture more likely. Finally, there is the rule of congruence. We are look¬ ing for activities that, in extant living organisms, are carried out by protein mole¬ cules, not by clays or other mineral surfaces. Barring authentic peptides of the pro¬ tein type, whose formation and faithful reproduction under prebiotic conditions are most unlikely, the multimers of my hypothetical mixture appear as the next-best molecules for building the kind of three-dimensional structures that determine enzyme catalysis. This does not by any means preclude the participation of metals and other cofactors in protometabolism. On the contrary. It is intriguing in this regard that a molecule like pantetheine could conceivably have been part of the multimer mixture.

Chapter 3

The Fuel of Emerging Life

Protometabolism

could not have unfolded without a supply of energy,

together with the means for productively exploiting this energy. The life-building complexification process was uphill all the way. To make it downhill and, there¬ fore, able to occur spontaneously, a sufficient supply of energy was essential. There were plenty of energy sources on the prebiotic Earth, in the form of sunlight, ultra¬ violet radiation, electric discharges, shock waves, heat, and chemical upheavals of various sorts. Which among these various sources of energy did emerging life exploit? And, especially, how was the raw power in the prebiotic setting converted into productive life-creating events?

THE PROBLEM OF PRIMEVAL MEMBRANES If, in accordance with the congruence rule, we ask present-day life for a hint with regard to these questions, we immediately run into a problem. The most important energy generators of living organisms today depend on the opeiation of highly complex substances organized within the fabric of an intricate, filmlike structure, or membrane. Could such arrangements have arisen early enough to satisfy the energy requirements of emerging life? Several authors believe so. In a book devoted to bioenergetics, Franklin Harold, a biochemist from the University of Colorado, does not hesitate to head an impoitant section with the declaration: “In the beginning was the membrane. 1 Clair Folsome, from the University of Hawaii, has proposed that primitive membranous vesicles might have formed from some oily “scum”—what 1 have called “goo

in

chapter 1—which must have been abundant in the prebiotic world, and that these vesicles might have developed into photochemical “protobionts” by association with some light-catching molecule.2 This and other similar proposals cannot be dis-

THE FUEL OF EMERGING LIFE

35

missed but seem to me incompatible with the necessarily rudimentary nature of the first energy-providing systems. Even if we assume the existence of some mem¬ brane-bound light-trapping system, we still have to account for the channeling of the trapped energy into productive chemical processes, rather than useless heat. Could incipient life have done without a membrane? Such is the question I shall ask in this chapter. As we shall see, there are good reasons for believing that life could, indeed, have done so. Properly interrogated, it even tells us how. A good introduction to the topic is provided by what may have been the first energy obsta¬ cle to be overcome in the development of life.

THE CASE OF THE MISSING HYDROGEN The problem is most easily defined in contemporary terms. As an example, take a bowl of spinach leaves and carefully dry them so as to remove all the water but no other volatile constituent. As every cook knows, not much will be left since spinach is “mostly water.” What is left, however, is what gave Popeye his superstrength. Give it to a chemist for elementary analysis and he will tell you that the material consists largely of carbon, oxygen, nitrogen, and hydrogen. In numbers of atoms, the proportions are roughly: C, 60; O, 40; N, 2; and H, 100. Consider now the nature of the “foodstuffs” with which the spinach plant makes its constituents. Car¬ bon comes from atmospheric carbon dioxide (C00), nitrogen from soil nitrate (N03“), and hydrogen from water (H0O). Now, try to make dry spinach from these building blocks and you find that you end up with a large surplus of oxygen: 120 atoms (60 x 2) brought in with carbon dioxide, 6 with nitrate (2 x 3), and 50 with water; a total of 176 oxygen atoms, or an excess of 136 atoms over the 40 that are needed. An alternative way of expressing this excess is in terms of the hydrogen that should be provided to convert the oxygen to water, that is, in the example con¬ sidered, 272 atoms of hydrogen, two for each excess atom of oxygen. In conclu¬ sion, contemporary autotrophic (self-constructing) life needs a source of hydrogen. How about emerging life? If the atmosphere had been as imagined by Urey, there would have been no prob¬ lem. Take the carbon from methane (CH4), the nitrogen from ammonia (NH3), and the oxygen from water, and you already have a large surplus of hydrogen (326 atoms against the 100 atoms needed), not counting the molecular hydrogen that Urey added for good measure. But with carbon dioxide as the source of carbon, you are in great trouble, as Miller found out experimentally. Even if Urey should finally turn out to be right, the evil moment would only be postponed. Sooner or later—probably sooner than later—hydrogen would be missing, as it is today. Where was it found? A naive sleuth, entrusted with the case of the missing hydrogen, might well answer: “What’s your problem? There was more hydrogen in the oceans than you

THE AGE OF CHEMISTRY

36

could ever use.” Only too true, except that it wasn’t there for the taking. Our sleuth forgot the golden rule that “you can’t have your cake and eat it,” which, in the chemical world, in the whole universe, in fact, translates to you can t nave it both ways.” You can’t take hydrogen from water and then use it to convert oxygen back to water. If you could, you would have invented perpetual motion, a dream pursued by generations of cranks, always in vain, because it is incompatible with one of the most profound laws of nature: For every natural event, there is an allowed direction and a forbidden direction. Apples fall; they don’t jump up to their branches. Sugar lumps dissolve in your coffee; they don’t form by themselves in a cup of sweetened coffee. Hydrogen combines with oxygen to form water; water does not dissociate spontaneously into a mixture of hydrogen and oxygen. All nature’s streets are one way. You can take them in the wrong direction, of course, but you have to work for it: lift the apple, extract the sugar, wrench the hydrogen off the water molecules, for example, with electricity. This is the absolutely fundamental way of nature, expressed by what scientists call the second law of thermodynamics, often shortened simply to the Second Law: If you have to work for it, you are going in the forbidden direction. In the allowed direction, on the other hand, the phenomenon, properly harnessed, may work for you, though it will never give you as much work as you would have to carry out yourself to reverse it. With a rope and pulley, you can use a falling apple to lift another apple, but only if the other apple is lighter. A convenient image to describe the two directions is downhill for what is allowed and uphill for what is forbidden. Perfectly horizontal means no work in either direction. It is the state of equilibrium, or perfect balance. Let us now return to the problem of the missing hydrogen and ask contemporary organisms how they solve it. Where does spinach get the extra hydrogen it needs for growth? The answer is: right where our sleuth pointed, from water. But, as demanded by the Second Law, the spinach has to work for it. Or, rather, it makes the sun work for it. Chlorophyll, the green stuff of plants, does exactly that. It uses the energy of sunlight to tear off hydrogen from water and lift it to a high enough level of energy so that the hydrogen can, in turn, tear off the oxygen from carbon dioxide and nitrate and replace it, doing everything under its own power, that is, going downhill. This concept of energy level is crucial. We can readily picture it by an image borrowed from gravity, although we are, of course, dealing with chemical energy. The higher a weight, the greater the amount of work that can be performed by its fall. In chemistry, height is replaced by other concepts, such as pressure, con¬ centration, potential, and suchlike. We shall do without these complications and adopt the less rigorous but readily grasped notion of energy level. Chlorophyll is a very complex molecule, which, in addition, has to be associated with other complex molecules embedded in a membrane to do its job. It is highly unlikely that such a system could have arisen spontaneously in early prebiotic days. We must search elsewhere for the prebiotic supply of hydrogen. To do this, we must first take a closer look at the hydrogen atom.

THE FUEL OF EMERGING LIFE

37

Hydrogen is the smallest of all atoms. It consists of a single, positively charged particle, or proton, which forms the nucleus of the atom and bears most of its mass, and of a single, peripheral electron, a negatively charged particle less than onethousandth the mass of thp proton. In the simplified picture of the atom known by the name of the great Danish physicist Niels Bohr, the electron is viewed as gravi¬ tating around the nucleus like a planet around the sun. Quantum mechanics pro¬ vides a more accurate but less intuitive image. The Bohr model of the atom will suffice for our purpose. As it happens, small quantities of free protons exist naturally in water as a result of the spontaneously occurring dissociation of the water molecule into positive hydrogen ions,TI+, which are no other than protons, and negative hydroxyl ions, OH-. In pure water, only one water molecule in ten million is thus dissociated (which still amounts to more than one million billion protons and hydroxyl ions in a teaspoonful of water). The number of free protons in water increases, and that of hydroxyl ions decreases, with increasing acidity (decreasing alkalinity), and vice versa. By definition, acids are substances that release protons in aqueous solution. Alkalis, or bases, are substances that pick up protons. This little excursion into physical chemistry was needed to clarify a point of fun¬ damental importance: It is possible to get hydrogen from water by supplying elec¬ trons. These can combine with protons, produced by the dissociation of water mol¬ ecules, to form hydrogen atoms. However, the Second Law should not be forgotten. If we wish the hydrogen to perform the job we have in mind—tear off oxygen from carbon dioxide and nitrate—we need the hydrogen to be at a sufficiently high level of energy so that it can go downhill henceforth. This implies that the electrons must themselves be supplied at a sufficiently high level of energy so that the hydrogen atoms they make by combining with protons be lifted to the energy level needed to tear off the oxygen from carbon dioxide and nitrate. In conclusion, in the presence of water to either absorb or provide protons, free hydrogen atoms and electrons are interchangeable. In biochemical jargon, the par¬ ticipation of protons is often made implicit, and one speaks simply of electrons. By definition, the gaining of electrons (or hydrogen) by a substance is called reduction; the loss of electrons (or hydrogen), oxidation. The two types of reactions are oblig¬ atorily linked. For one substance to be reduced, another has to be oxidized so as to provide the necessary electrons (or hydrogen atoms). Thus, we are always dealing with oxidation-reduction reactions, or, as more commonly termed nowadays, elec¬ tron-transfer reactions. The notion of energy level must be kept in mind. When electrons are transferred, the electron donor is, by definition, the substance in which transferable electrons occupy the higher energy level, and the electron acceptor is the substance that, when reduced, has electrons occupying a lower energy level. Electrons move from donor to acceptor in the downhill direction, like every other happening in the world. Equipped with this information, we may now search the prebiotic setting for a suitable source of electrons to effect the required reactions, which will henceforth

38

THE AGE OF CHEMISTRY

be referred to as biosynthetic reductions. Several solutions to this problem have been proposed. I shall mention only two, which happen both to involve iron. The first mechanism uses sunlight energy to remove hydrogen from water, as does the plant system, but has the immense advantage that it needs no intricate catalyst to operate. The reaction takes place in simple aqueous solution, its only requirement being the presence of iron atoms in the form of ferrous ions (Fe2+), which carry two positive charges.3 (Ions are electrically charged atoms or molecules.) We have seen that the prebiotic oceans contained considerable amounts of this material. The source of energy is ultraviolet (UV) light, rather than visible light, but this poses no problem since the prebiotic Earth was exposed to strong UV radiation. When a fer¬ rous ion is excited by a photon of UV light, it relinquishes an electron and changes to the ferric form (Fe3+), which carries three positive charges. The electron com¬ bines with a proton to give rise to a hydrogen atom. In this process, electrons are transferred from ferrous iron (the donor) to protons (the acceptor). In the reverse reaction, hydrogen would be the donor and ferric iron the acceptor. In the absence of UV light, the downhill direction of the transfer would actually be the latter. Thanks to the energy provided by the UV light, the spontaneous direction of the reaction is reversed and the electrons released by ferrous iron are lifted to a conve¬ niently high energy level that allows them to serve in prebiotic reductions. A possible sign that such a reaction occurred is found in deposits of the mineral magnetite, a mixed oxide of ferrous and ferric iron found in iron-rich geological strata called banded iron-formations4 because of their striped appearance. The age of banded iron-formations ranges from 1.5 to over 3.5 billion years. It is usually accepted that these formations arose from the interaction of ferrous iron with the oxygen produced by light-utilizing bacteria, but the possibility that the Unsup¬ ported reaction just described contributed to their creation is not to be excluded. Another possible source of prebiotic electrons is hydrogen sulfide, a typical part of the prebiotic setting. Wachtershauser has proposed a reaction in which, in the presence of ferrous iron, two sulfide ions (SH~)—which form from hydrogen sul¬ fide in aqueous solution—would be converted into a disulfide ion (S?2~) with the release of molecular hydrogen. In this case, the iron does not abandon an electron. Instead, it drives the reaction by combining with the generated disulfide to form the highly insoluble ferrous disulfide (FeS2), thereby removing the product of the re¬ action from solution and allowing more to be formed. The validity of this model has been proved experimentally.5 Ferrous disulfide is the constituent of pyrite, the mineral assumed to provide a catalytic surface in Wachtershauser’s protometabolic model discussed in the preceding chapter. We thus have two model systems that could theoretically account for the missing hydrogen. They are not mutually exclusive. The two reactions could have occurred side by side or in different environments. By definition, the UV-supported reaction could have taken place only in surface water layers, whereas the pyrite-generating process could have occurred in the dark depths of the ocean. In conclusion, whatever the exact composition of the prebiotic atmosphere, we

THE FUEL OF EMERGING LIFE

39

may consider it likely that our young planet, abundantly supplied with ferrous iron, enveloped in hydrogen sulfide fumes, and exposed to strong ultraviolet radiation, was exuding hydrogen through every pore, while iron combinations destined one day to become the minerals magnetite and pyrite were laid down at the bottom of the oceans. Electrons were indeed available at a high enough energy level to sup¬ port early biosynthetic reductions. Interestingly, iron and sulfur are both key constituents of catalysts that partici¬ pate in electron-transfer reactions in present-day living organisms. The most ancient such catalysts may well be proteins called iron-sulfur proteins,6 in which the catalytic center is an iron atom, which oscillates between the ferrous and the ferric state, surrounded by a sulfide cluster. The hint is unmistakable.

THE CASE OF THE EXCESS WATER A good supply of electrons solves only half of the prebiotic energy problem. The other half concerns the joining of molecules in a watery medium. Examples of such reactions already encountered are the formation of esters from alcohols and acids, of thioesters from thiols and acids, of peptides from amino acids, of nucleotides from phosphate, ribose, and a base, and of RNA from nucleotides. Many more such condensation reactions occur in living organisms. All have in common that they take place with the removal of water. In aqueous solution, this is the forbidden direction, as was explained in chapter 2. Nature’s universal answer to the problem is ATP. Almost as famous as DNA, this acronym stands for adenosine triphosphate. Adenosine is the combination of the purine base adenine with ribose. In association with one phosphate molecule, adenosine forms adenosine monophosphate (AMP), one of the four nucleotides that serve to make RNA. With a second phosphate attached to the phosphate of AMP, one gets adenosine diphosphate (ADP), which, with an additional phosphate attached to its terminal phosphate, gives ATP. The two bonds linking the three phosphates of ATP are called pyrophosphate bonds, after inorganic pyrophosphate (PPj), a combination of two phosphates that arises when inorganic phosphate (P.) is heated at a high temperature (pyr means fire in Greek). The formation of pyrophosphate bonds is a typical condensation reac¬ tion. It takes place with the removal of a water molecule. In the reverse reaction, the bonds are split with the help of water, by hydrolysis. As for all such bonds in an aqueous milieu, hydrolysis goes downhill, condensation uphill. ATP is the universal biological condensing agent, that is, chemical water extrac¬ tor. Its hydrolysis, either to ADP and P| or to AMP and PR, serves to extract the water that needs to be removed in order to seal the created bond. This takes place by special mechanisms known as sequential group transfers that act in such a way that the water molecule involved is transferred directly from the joining molecules

40

THE AGE OF CHEMISTRY

that generate it to the molecule (ATP) that consumes it for hydrolytic splitting; it never appears in free form nor mixes with the surrounding water. In such a transfer, the water molecule takes the direction of least resistance (downhill). If less energy is required to seal a bond between X and Y than between ADP and P or between AMP and PP.—that is, if ATP hydrolysis releases more energy than is needed for joining X and Y—X-Y will be formed and a pyrophosphate bond of ATP will be split. The reverse will occur if the opposite is true. If the formation of the two bonds requires equal amounts of energy—the reaction is freely reversible—the exchange will be partial according to the rules of chemical equilibrium. As it happens, most of the bonds in biological substances require less energy for their formation than do the pyrophosphate bonds of ATP, which explains why ATP is an efficacious condensing agent. The biochemist Fritz Lipmann has named the pyrophosphate bonds of ATP high-energy bonds for this reason.7 The bonds in pro¬ teins and other natural substances that are sealed at the expense of ATP hydrolysis are low-energy bonds. ATP is not the ultimate source of energy for condensation reactions. We don’t get ATP from our food, and our cells contain this vital substance in only small amounts. Life would quickly grind to a standstill were not ATP continually reassembled from its hydrolysis products. This problem will be examined later in this chapter. First, let us ask whether a substance such as ATP could have been available to emerging life. Surely not ATP itself. It is too complex a molecule for this early phase of prebiotic events. With ATP, we are well into the RNA world, not at the onset of protome¬ tabolism. But what about the simpler inorganic pyrophosphate molecule? The pyrophosphate bond of inorganic pyrophosphate is not as energetically powerful as the similar bonds of ATP, but it is sufficiently energy-rich to substitute for the ATP bonds in many processes. There is ample evidence in the present living world that inorganic pyrophosphate can carry out the same basic functions as ATP. Most researchers believe that pyrophosphate preceded ATP as the first bearer of productive high-energy bonds.8 This role could also have been played by polyphos¬ phates, associations of a larger number of phosphate groups linked by pyrophos¬ phate bonds, which are also found in a number of organisms. Accordingly, many scientists have searched the geological record for the possible prebiotic occurrence of such substances. The results of this search are not promising. I mentioned in chapter 1 the rarity of soluble inorganic phosphate and the problem this poses in relation to the overwhelm¬ ing biological importance of phosphate. The problem is compounded in the case of pyrophosphates and polyphosphates, which are much less abundant than phosphates and also locked up in water-insoluble combinations. However, acid, which I sug¬ gested as a possible means of solubilizing phosphate, could have done the same for pyrophosphate. Also, the volcanic production of pyrophosphate has been detected recently and could have been a more abundant source under prebiotic conditions.9 An alternative possibility is that thioesters were the primeval source of energy for

THE FUEL OF EMERGING LIFE

41

condensation reactions.10 Thioesters are obligatory intermediates in a large number of ATP-supported condensation reactions in which one of the partners of the reaction is an acid. In such reactions, the step powered by ATP hydrolysis is the condensation of the acid with a thiol, usually pantetheine phosphate or coenzyme A, to form the corresponding thioester. The acid group is then transferred from the thioester to its acceptor. We saw in chapter 2 how such group-transfer reactions are involved in the formation of esters and some peptides. Many other important biological constituents are made by way of thioesters, including a large variety of fatty substances, choles¬ terol, several vitamins, parts of chlorophyll, and numerous metabolic intermediates. What makes thioesters attractive is that they are energetically equivalent to ATP. The thioester bond is a high-energy bond. Thus, thioesters can support the assembly of ATP or be assembled at the expense of ATP hydrolysis equally well. The Ameri¬ can chemist Arthur Weber, formerly of the Salk Institute in San Diego, now at the NASA Ames Research Center in Moffett Field, California, who has pioneered the study of prebiotic sulfur compounds, has shown that thioesters can, under very sim¬ ple conditions, support the formation of inorganic pyrophosphate from inorganic phosphate by a mechanism similar to that of the present-day thioester-linked process whereby ADP and P; are joined into ATP.11 We thus have the choice between two possibilities consistent with the congru¬ ence rule. Pyrophosphate was provided by the prebiotic setting and served as a con¬ densing agent for the assembly of thioesters. Or, conversely, thioesters came first and supported the assembly of pyrophosphate. Or, of course, the two could have appeared independently and interacted later. Before we come to any conclusion, we must look first at the mechanisms whereby ATP is continually regenerated from its hydrolysis products in extant living organisms.

HOW THE WHEELS ARE KEPT TURNING In living cells, ATP turns over very rapidly. It is continually consumed—that is, split hydrolytically—by the performance of chemical work (and many other kinds of work, as we shall see later), and it is regenerated equally fast from its hydrolysis products. Where does the energy required for this regeneration process come from? The answer to that vital question brings us back to the case of the missing hydro¬ gen. The energy for the regeneration of ATP comes from electron flow. We saw in the beginning of this chapter how electrons can be transferred from a reduced donor occupying a higher energy level to an oxidized acceptor occupying a lower energy level. A transfer of this sort releases energy in an amount proportional (per electron transferred) to the difference in the two energy levels. As a simple image, think of a waterfall. The energy released by the fall of a given amount of water is proportional to the height of the fall.

42

THE AGE OF CHEMISTRY

In all living cells, certain key “electron falls” are coupled to the assembly of ATP from ADP and P., the way certain waterfalls are harnessed to the running of a mill or to the generation of electricity. This universal mechanism is called oxidative phosphorylation—oxidative because the electron donor in the coupled reaction is oxidized; phosphorylation because ADP is phosphorylated, that is, fitted with an additional phosphate group in the process. It requires three conditions: (1) an appropriate source of electrons; (2) an outlet for the electrons situated at a suffi¬ ciently lower energy level for the amount of energy released by the electron transfer to cover the needs of ATP assembly (as a rule, one ATP molecule is assembled for each pair of electrons transferred); and (3) a coupling system—the equivalent of the waterwheel or turbine in the waterfall analogy—linking ATP assembly to electron flow. Many different electron donors and acceptors are used in such reactions in nature. In organisms like ourselves, for example, the foodstuffs provide the elec¬ trons, and oxygen is the final acceptor. This is what really happens when we “burn” our food. Thanks to this mechanism, the energy released by such combustions is only partly given off as heat. Much of it is retrieved in the form of reassembled ATP. In illuminated green plants, the electrons are delivered at a high energy level by excited chlorophyll molecules and recovered at the lower level by the same mol¬ ecules. In between, the electrons fall through coupled phosphorylation systems, as they do in our tissues. Electron donors and acceptors vary, but the energy-retrieval mechanisms are universal. The most important such mechanisms are situated in membranes. We shall meet them later when we look at the first cells. We have agreed not to take them into consideration at this early stage in the origin of life. Some coupled phosphoryla¬ tions do not depend on membranes and take place in the cell sap, the soluble part of living cells. Known technically as substrate-level phosphorylations, these mecha¬ nisms could qualify as prebiotic (with pyrophosphate being assembled instead of ATP). Interestingly, they involve thioesters as key intermediates. The reaction directly coupled to the energy-releasing electron transfer is the formation of a thioester, which then supports ATP assembly in the manner already seen. Thioesters thus occupy a unique position in metabolism: They bridge the two main forms of biological energy—one linked to electron transfer and one linked to group transfer. In addition, we saw in the preceding chapter that thioesters could have played a key role in the generation of the first catalysts of emerging life. These facts, taken together with the probable abundance of acids and thiols on the prebi¬ otic Earth, build a strong case in favor of thioesters being the primary energizers of the early biogenic process, perhaps preceding inorganic pyrophosphate. But there is an if—and oh what a big if, to quote Darwin!12 The assembly of the primeval thioesters would itself have required energy. So, we are back to square one. There are several possible solutions to the problem of thioester assembly under prebiotic conditions. According to thermodynamic data, thioesters could form spontaneously from free acids and thiols in a watery medium, if this medium were

THE FUEL OF EMERGING LIFE

43

very hot and acidic. Even so, however, the yields would still be slight. Neverthe¬ less, the possibility is worth entertaining. Boiling acid may not be our idea of a cozy little niche. Neither is it a particularly favorable medium for a number of frag¬ ile biological molecules. Yet it is the choice habitat of certain bacteria known as thermoacidophilic, of particularly ancient origin.13 A number of authors believe that life started in a hot environment. They have paid particular attention in this connec¬ tion to the continuing recycling of water through deep-sea hydrothermal vents. One could imagine a scenario in which thioesters arising in the hot, acidic, sulfurous depths would be continually delivered to the prebiotic soup, where milder condi¬ tions prevailed and the energy-rich thioesters could carry out their function. This is not the only possibility. Weber has described other plausible mechanisms for the formation of thioesters.14 There is also the possibility, as yet unexplored, that thioesters may have assembled in the atmosphere from volatile thiols and acids. Finally, and perhaps most simply, there is the recourse to coupled electron transfer as a source of energy, as occurs in metabolism today. What is known of this reaction indicates that it could have taken place under prebiotic conditions. The necessary ingredients were probably present, and primi¬ tive iron-sulfur complexes, ancestral to iron-sulfur proteins, could have catalyzed such a reaction. In some ancient bacteria, a process of this sort is indeed catalyzed by an iron-sulfur protein.15 As to the electron acceptor required by the thioester-generating process, an attractive possibility is ferric iron, the product of the reaction whereby ferrous iron serves to generate hydrogen from protons with the help of UV light. By serving as electron acceptor, ferric iron would return to the ferrous state, thus completing a cycle in which electrons are released from ferrous iron with the help of UV light and return to ferrous iron by way of a complex pathway completed by the sealing of thioester bonds. The net result would be the use of UV-light energy to close thioester bonds, which, in turn, could serve to meet all the energy needs of emerg¬ ing life. Such a cycle would be entirely analogous to the water-oxygen cycle that supports much of the present-day biosphere—plants release oxygen from water with the help of visible light, and animals and other aerobic organisms use the oxy¬ gen as final electron acceptor and reconvert it into water—with the crucial differ¬ ence that the water-oxygen cycle requires complex structures, whereas this iron cycle does not. The collaboration of iron and sulfur in such a cycle could have been a primeval manifestation of the contemporary fecund alliance between these two biologically important elements.16

THE THIOESTER WORLD We have not uncovered the hidden trail of protometabolism, but we have found some telltale traces. These traces have been described in detail in this and the pre-

THE AGE OF CHEMISTRY

44

ceding chapter. A brief summary of what they have revealed may be useful. The main message comes out loud and clear: sulfur. This element is, quantitatively, a minor component of living matter, but qualita¬ tively a very important one. Two of the twenty amino acids that form proteins, cys¬ teine and methionine, contain sulfur. So do several coenzymes. Sulfur is often found at the heart of the catalytic centers of enzymes. There is sulfur also in a num¬ ber of structural macromolecules, for example, some of the main components of cartilage. Many of the most ancient bacteria live by metabolizing some sulfur com¬ pound. The prebiotic world was steeped in sulfur. It all amounts to a very strong case. In present-day organisms, sulfur is derived mostly from the fully oxidized sul¬ fate ion (S042~), which exists unchanged in a number of components, mostly struc¬ tural, in which its role is mainly to provide negative charges to the molecules. Many of the most important biological functions of sulfur, however, require sulfate to be reduced to hydrogen sulfide (H0S) and incorporated into organic molecules, mostly thiols and their derivatives. Hydrogen sulfide is also the form of sulfur that dominated in the prebiotic world. The traces we have detected unmistakably point to thiols. In the prebiotic soup, thiols most likely were present together with a variety of amino acids and other organic acids, which are the main substances produced in Miller-type simulation experiments and found in meteorites. Thiols and acids read¬ ily join into thioesters, provided some means exist for the removal of the water molecule that must be extracted for a thioester bond to form. Several possible mechanisms whereby this could have happened exist. My main hypothesis is that, somewhere in the prebiotic world, conditions prevailed under which thioesters formed spontaneously. Granted this unproved but not implausible assumption, the way opens to a metabolism-like protometabolism supported by thioesters. Thioesters provided protometabolism with two essential ingredients: catalysis and energy. The catalysts were peptides and peptide-like substances that prefigured present-day enzymes and guided the first building blocks of life into directions not too different from present-day metabolism. The energy was in a form that fitted within such pathways and could have served to usher in inorganic phosphate and the all-important pyrophosphate bond. The high-grade electrons required for the first biosynthetic reductions could have been provided by ferrous iron with the help of UV light or by hydrogen sul¬ fide with the help of ferrous iron. The first mechanism would have given rise to fer¬ ric ions, which could have served as the first electron acceptor in energy-yielding electron-transfer reactions coupled to the synthesis of thioesters and, in due course, of inorganic pyrophosphate. Together, the two processes would have closed an iron cycle whereby UV-light energy supported the assembly of thioesters and, by way of the splitting of the thioesters, the whole of protometabolism. In addition, iron asso¬ ciated with sulfide could have constituted the first electron-transfer catalyst. This “thioester world,” or, better said, “thioester-iron world,” represents my

THE FUEL OF EMERGING LIFE

45

hypothetical reconstruction, based on the few traces it has left, of the hidden trail that led from the first products of prebiotic chemistry to the RNA world, and that continued to support the RNA world during all the time it took emerging life to evolve from the RNA world to the RNA-protein world. This view of the trail is purely conjectural. Quite possibly, future findings will point to different pathways not thought of today. For my part, I would find it very suprising if these early path¬ ways did not reveal glimpses of present-day metabolism.

✓

Chapter U

The Advent of RNA

Even if we accept the premises of a thioester world, we are no nearer to identi¬ fying the chemical pathways that led from the first building blocks of life to RNA. A possible experimental approach to the problem does suggest itself: Reproduce in the laboratory the primeval mixture of multimers and look for key catalytic activi¬ ties in the mixture. According to my model, this is the heart of the problem. Pro¬ tometabolism must have been channeled by the early catalysts the way metabolism is by enzymes today. Several techniques now exist for making random peptides. One could even go back to the old Wieland procedure, which has the merit of actually using thioesters. Such experiments might not provide the selective conditions that led to the particu¬ lar subset of multimers assumed by the model, but they would be a step in the right direction. The congruence rule, on the other hand, would help in choosing the kind of catalytic activities for which to search. I am, unfortunately, too far along my own trail to start such an approach. But other laboratories are becoming interested in it. Meanwhile, we are left to conjecture, using present-day metabolism as a guide. A possible clue is provided by ATP.

THE ATP CONNECTION ATP plays a key role in energy metabolism. It is also one of the four precursor mol¬ ecules used in the synthesis of RNA. Here lies the connection. RNA molecules are constructed from nucleotides, which are combinations of phosphate, ribose, and one of four bases: adenine, guanine, cytosine, and uracil. AMP, the parent molecule of ATP, is one such nucleotide. The similarly constructed GMP, CMP, and UMP are the other three. Just as AMP can be phosphorylated to ADP and ATP, the other nucleotides can likewise be converted to GDP and GTP, CDP and CTP, and UDP and UTP, respectively. The pyrophosphate bonds in GTP, CTP, and UTP have the

THE ADVENT OF R N A

47

same properties as the pyrophosphate bonds in ATP; their splitting can similarly support energy-requiring processes and does so in certain cases. Even when this happens, the central role remains the prerogative of ATP. Whenever another energy donor is involved, it is regenerated at the expense of ATP splitting. Why was ATP singled out in this manner? A possible answer to this question is that adenine happened to be there before the other bases. It is certainly the easiest one to make abiotically. Oro’s celebrated synthesis of adenine from ammonium cyanide was mentioned in chapter 2.1 Even though the relevance of this finding has not been established, it suggests that adenine may belong to the group of easily formed molecules that made up the early building blocks of life. In support of this hypothesis, trades of adenine have been detected on meteorites.2 The origin of the 5-carbon sugar ribose is obscure. Sugars arise readily from formaldehyde in alkaline solution, but as a complex mixture of different molecules. In metabolism, ribose is formed from the 6-carbon sugar glucose by a devious path¬ way that involves phosphate-linked intermediates. Work by the Swiss chemist Albert Eschenmoser, from the Federal Institute of Technology in Zurich, has shown that phosphate groups may, even in the absence of a catalyst, exert highly selective influences on the reactivity of sugar molecules.3 Phosphate groups are also involved in the biological mechanism whereby ribose joins with bases, and they provide the phosphate component of nucleotides. Perhaps the early appearance of inorganic pyrophosphate—from natural sources or by way of the thioesters of my model—helped steer protometabolism in the direction of AMP formation. At pres¬ ent, one can only surmise. Once AMP comes on the scene, an interesting possibility is offered by the thioester-world model. It is known that the thioester bond and the pyrophosphate bond are energetically equivalent; the hydrolysis of either one can support the water-releasing assembly of the other. In one such reaction, ATP hydrolysis to AMP and inorganic pyrophosphate (PR) serves in the condensation of acids with co¬ enzyme A or pantetheine phosphate (two major thiol cofactors in metabolism). This is how the thioesters involved in the biosynthesis of esters and many other impor¬ tant biological compounds are assembled. The reaction is freely reversible. Now, imagine yourself in a thioester world containing thioesters and PP., at the time when the first molecules of AMP made their appearance. Through a reversal of the reaction just mentioned, AMP could have joined with PP with the support of the energy supplied by thioester hydrolysis.4 Had you been there, you would have witnessed one of the great events in the chemical origin of life: the birth of ATP, the universal purveyor of biological energy, destined eventually to replace in all its major functions the pyrophosphate from which it arose. An interesting possibility is that ATP, in turn, served to usher in RNA: in other words, information may have entered by way of energy. What may have happened is the condensation of two ATP molecules to ATP-AMP, with the release of inor¬ ganic pyrophosphate. In this reaction, the second ATP donates an AMP molecule to the first ATP, and the bond between ATP and AMP is sealed at the expense of the

48

THE AGE OF CHEMISTRY

bond that originally existed between AMP and pyrophosphate in the AMP-donating ATP molecule. Have another ATP donate AMP to ATP-AMP, and you get ATPAMP-AMP Repeat the reaction any number of times, and you get a chain of ATPAMP-AMP-AMP— of any length, called poly-A. This is not science fiction. Such a reaction actually occurs in many living cells, where it adds poly-A tails up to about 250 nucleotides long to many RNA molecules. Unlike the synthesis of true RNA, the formation of poly-A takes place without the supply of information; it consists merely of “dumb,” repetitive assembly. In the context of protometabolism, poly-A would have been no more than a stor¬ age form of AMP molecules, which would automatically have regulated the relative amounts of ATP and PPj present. Imagine a situation where ATP was abundant and PPj scarce. Because of the laws of chemical equilibrium, ATP would be driven to form poly-A, and PP; would rise. In the opposite situation, excess PP; would drive the formation of ATP from poly-A. The process may have been “dumb,” as was the whole of protometabolism before RNA appeared. But the outcome was far from useless. It preserved AMP and adjusted the availability of the two forms of highenergy pyrophosphate bonds—PPj and ATP—to their relative rates of consumption. It is not known how the other three bases found in RNA appeared. Possible schemes for their abiotic formation have been suggested.5 An origin from adenine is not inconceivable. Once they were present, their incorporation into nucleotides could have been brought about by AMP. Reactions are known by which one base replaces another in nucleotides. Thus, guanine could replace adenine in AMP, to form GMP; cytosine could similarly lead to CMP, uracil to UMP. In turn, these nucleotides could acquire phosphate groups from ATP, as they do in present-day metabolism, and give rise to GTP, CTP, and UTP. Finally, the same “dumb” reac¬ tion that caused poly-A to be formed from ATP could have incorporated GMP, CMP, and UMP into similar associations, which would have been the first RNA molecules, though still devoid of information, a mere jumble of letters. This sce¬ nario is conjectural but not unlikely. It seems reasonable to suppose that the chem¬ istry came first, without information, and that information came later. In the present-day world, RNA is assembled as described from ATP, GTP, CTP, and UTP, which donate AMP, GMP, CMP, and UMP units to a growing chain initi¬ ated by ATP or GTP. What the biological process has in addition, and the prebiotic process may have lacked, is a mechanism for the selection of whichever of the four available nucleotides is to be added at each step. The birth of this mechanism, which relies on simple molecular interactions, represents a true watershed in the development of life on Earth. It signals the transition from the age of chemistry to the age of information. Before we ourselves effect this transition, we must take a look at an important group of catalytic components known as coenzymes, many of which may go back to the age of chemistry. Their existence raises interesting possibilities with respect to the functioning of the RNA world and the origin of RNA.

THE ADVENT OF RN A

49

COENZYMES: CHILDREN OF THE R N A WORLD? In metabolism, enzymes are often assisted by special molecules called coenzymes. These serve most frequently as intermediaries, or carriers, in transfer reactions. Imagine a process in which entity X is transferred from the X donor X-Y to the acceptor Z, to give X-Z and Y. In many cases, a carrier K mediates the transfer so that X is first transferred from X-Y to K, with the formation of X-K, which then donates X to Z, making X-Z and leaving K ready for another round. There are sev¬ eral advantages to this peculiarity. A major advantage is centralization and simplifi¬ cation of the transfers of the same entity. Suppose, for example, that X is to be exchanged between ten different donors and ten different acceptors. The number of individual reactions needed to allow all possible exchanges is one hundred. With K as common intermediary, only twenty reactions are required. It is like having a cen¬ tral currency, as against barter. Two kinds of entities are exchanged in transfer reactions: electrons or chemical groups. We have seen the importance of electron transfers in biosynthetic reduc¬ tions and in energy retrieval. As to group transfers, they represent the main mecha¬ nism of biosynthetic assembly reactions. Several examples were mentioned in chapters 2 and 3. RNA synthesis is another. In RNA elongation, the addition of, say, AMP is a transfer of the AMP group from ATP to the growing chain. Virtually every biological condensation proceeds by group transfer. Attesting to the impor¬ tance of transfer reactions is the fact that more than 90 percent of the reactions that occur in living organisms are either electron transfers or group transfers. Most of these reactions take place with the help of a coenzyme carrier. There are thus two main categories of coenzymes: electron carriers and group carriers. Some coenzymes may go back to early prebiotic days. We have already encoun¬ tered iron-sulfur complexes as putative primitive electron carriers. Some thiols, such as coenzyme M or pantetheine phosphate, may be among the earliest group carriers. Many other coenzymes could be children of the RNA world. It is striking that all four nucleotides provide important group carriers participating in the syn¬ thesis of certain carbohydrate (derived from sugars) and lipid (fats) components. In addition, AMP is part of several other coenzymes, including coenzyme A, where it is linked with pantetheine phosphate, and several central electron carriers. In a number of coenzymes, the active part is represented by a flat, ring-shaped, nitrogenous molecule chemically related to the bases found in the major nucleotides. Many of these special molecules are vitamins, that is, essential chemi¬ cals that the human organism is incapable of making and that must be supplied with food. Interestingly, some of these substances are engaged in combinations entirely similar to nucleotides. Such is the case of nicotinamide, an important vitamin known as vitamin PP (pellagra preventiva), whose deficiency causes pellagra, a

THE AGE OF CHEMISTRY

50

severe nutritional disease that was once prevalent—and still is in remote areas—in many parts of Latin America. In living organisms, nicotinamide is linked with ribose and phosphate in a typical nucleotide, nicotinamide mononucleotide, or NMN. In association with AMP, NMN forms the two major electron carriers, NAD (nicotinamide adenine dinucleotide) and NADP (nicotinamide adenine dinucleotide phosphate). Another vitamin, riboflavin, or vitamin B^, is similarly engaged in a nucleotide¬ like combination (ribose is replaced by a related substance), flavin mononucleotide (FMN), which, combined with AMP, gives flavin adenine dinucleotide (FAD). Both FMN and FAD are also important electron carriers. The fact that so many coenzymes are nucleotides is often quoted in support of the RNA-world model. These molecules are seen as going back to an RNA world entirely operated by ribozymes intimately connected with nucleotide coenzymes. In the words of the American biochemist Harold White III, the nucleotide coenzymes could be “fossils of an earlier metabolic state.”6 The existence of molecules such as NMN and FMN raises the possibility that the first RNA molecules contained more than four kinds of nucleotides. The condi¬ tions that led to the abiotic synthesis of purines (adenine and guanine) and pyrim¬ idines (cytosine and uracil) may well have spawned the formation of a whole array of related nitrogenous bases. A number of these could have formed nucleotides, which, in turn, could have become incorporated into RNA-like combinations. The four RNA components would then have been selected later by virtue of their unique ability to support information transfer. This intriguing possibility will be considered in the following chapter.

THE CHEMICAL FOUNDATION OF LIFE The most important notion encountered in our attempt to reconstruct the early path¬ way from abiotic chemistry to the RNA world is that of congruence between pro¬ tometabolism and metabolism. This notion runs contrary to the ideas generally accepted in the field. Here is how two pioneers in origin-of-life research, Miller and Orgel, summed up their views on the topic in 1973. Referring to the possibility that “metabolic pathways parallel the corresponding prebiotic syntheses that occurred on the primitive earth,” they wrote: “It is not difficult to show that this hypothesis cannot be correct in the majority of cases. Perhaps the strongest evidence comes from a direct comparison of known contemporary biosynthetic pathways with rea¬ sonable prebiotic pathways—in general they do not correspond at all.”7 What is a “reasonable” prebiotic pathway? No doubt, the very early reactions, such as we know have occurred on comets and meteorites and suspect occurred on Earth, are products of some basic reactions of organic chemistry, of the kind Miller

THE ADVENT OF R N A

51

calls “robust,”8 that require only simple conditions to take place. A number of amino acids and other organic acids, perhaps adenine and some other nitrogenous bases, plausibly some sugars (although this is more problematical), could have formed in this way. These substances are what I call the building blocks of life. However, the road from these simple molecules to the kind of chemical com¬ plexity required to generate and sustain an RNA world belongs to a different realm. There is no robust way of making ATP, let alone an RNA molecule. This, in itself, makes a strong argument in support of a catalyst-mediated protometabolism. A further conclusion, which I believe to be compelling, is that protometabolism must have prefigured metabolism. I don’t see how the RNA world could, by the progressive generation of protein enzymes, have given rise to a network of chemi¬ cal reactions unrelated to those that produced and sustained it in the first place. Metabolism must have arisen congruently with protometabolism. Thus, the basic chemistry that underlies life in all its forms was laid down right from the start, by a succession of steps ruled by the strictly deterministic factors that govern all chemical processes. This is true not only of the particular model I have proposed, but of any model. Whatever pathways emerging life followed on its way to the RNA world, the nature of the process as a multistep succession of chem¬ ical events precludes a significant participation of improbable accidents. The bio¬ genic pathway involved reactions that were bound to take place under prevailing conditions. Another conclusion is that the early phase of the biogenic process must have been fast, contrary to the commonly accepted notion that the emergence of life took a very long time.9 With the kind of fragile chemicals involved in the construction of life, only a fast process could overcome the wear of spontaneous decay. In a primeval soup appropriately stocked with building blocks, sources of energy, and catalysts, RNA could be reached in a matter of years, if not less. A large number of such developments—many abortive for one reason or the other—could have been initiated in different parts of the world and at different times. Even the possibility of one day reproducing the process in the laboratory no longer belongs to the realm of science fiction.

PART II

THE AGE OF INFORMATION

Chapter 5

RNA Takes Over

Little has been said so far about information, the reason being that what¬

ever led to the appearance of the first RNA molecules, it was not an anticipation of their informational role. Initially, RNA was a product of chemical determinism. Information arose as an emergent property. A few words concerning this property will be helpful before we proceed.

A GLIMPSE OF THINGS TO COME All living organisms are constructed according to a blueprint that is transmitted from generation to generation. Plague bacilli beget plague bacilli, orchids beget orchids, mites beget mites, humans beget humans. For this reason, the blueprint is called genetic (the root gen, as in genesis, comes from the Greek verb meaning “to be bom”). The genetic blueprint consists of units, or genes, which together make up the genome, or genotype, of the organism. Genes have two cardinal properties: (1) they can be copied, the condition of hereditary transmission; and (2) they can be expressed into properties of the organism, which, together, constitute its phenotype. The underlying phenomena are entirely chemical. In all extant forms of life, the genetic blueprint is written into molecules of deoxyribonucleic acid (DNA), a material closely related to RNA and similarly con¬ structed from four different kinds of nucleotides. The sequence of nucleotides determines the informational content of the molecules—just as the sequence of let¬ ters determines the informational content of words. DNA is the stuff of our genes and, for this reason, deserves its preeminent posi¬ tion in the symbolism of life. Its function, however, is strictly limited to the storage of genetic information (and the replication of that information when a cell divides, so that each daughter cell has a copy). When it comes to expression of the informa¬ tion, DNA is invariably transcribed first into RNA. Transcription is not very differ-

THE AGE OF INFORMATION

56

ent from replication, as RNA is chemically similar to DNA and is constructed, but for a single, minor difference, with the same four bases. RNA is a much more versatile molecule than DNA. It can exhibit catalytic activ¬ ities, as in ribozymes, and thereby expresses the information received from the transcribed DNA in the performance of chemical reactions. Among these reactions are changes affecting RNA molecules and—within the framework of a complex cell structure, called a ribosome, made of a number of RNA and protein mole¬ cules—the assembly of amino acids into proteins, a process of overwhelming importance in all living beings. Such functional RNAs express only a small number of DNA genes. Most genes code for proteins, which, through their structural, regulatory, and especially cat¬ alytic functions (enzymes), are the main agents of phenotypic expression. Cells, by and large, and the organisms they serve to build, are the expression of their pro¬ teins. The sequences of amino acids in proteins are specified by the sequences of nucleotides in the DNA genes, though not directly but by way of RNA transcripts, which are the information “messengers” in this process. Because proteins are con¬ structed with an “alphabet” of twenty amino acids, as opposed to the fournucleotide alphabet of RNA (or DNA), the transfer of information from RNA to protein is called translation. The set of equivalences that govern translation forms the genetic code. These relationships are summarized by the sequence:

Protein (translation) (,replication) in which the arrows represent information transfers. Crick’s Central Dogma1 amounts to the amply verified statement that the last arrow (translation) is strictly unidirectional—reverse translation does not occur. This fact, together with the important catalytic role of certain RNA molecules in protein synthesis, explains why most investigators believe that RNA came before protein. But why RNA first, not DNA first? The reason for this is that DNA is theoretically expendable, whereas RNA is not. All that is needed is for RNA to be replicatable:

Protein (translation) (.replication)

RNA TAKES OVER

57

Such a process does not occur in normal cells, but it takes place in cells infected by certain viruses—the agent of polio, for example—that have an RNA genome. In these viruses, RNA is the replicatable storage form of genetic information; DNA is not involved. Among the,proteins coded for by the viral RNA is RNA replicase, the enzyme catalyzing RNA replication. Thus briefed, we may now attempt to reconstruct the historical events that led to the emergence of replication, protein synthesis, and translation, remembering that these developments came about through strictly chemical processes. Information did not beckon. It slipped in. As soon as it had done so, it became the central dri¬ ving force of life, by making possible Darwinian selection.

THE MAGIC CIPHER The first RNA molecules were probably random associations of nucleotides, which may, quite possibly, have contained other bases besides adenine, guanine, cytosine, and uracil. These bases are customarily represented by their initials. (In biochemi¬ cal abbreviations, such as ATP or UMP, A, G, C, and U stand for the combinations of the bases with ribose; but in the informational—and informal—shorthand of molecular biology, these chemical nuances are dispensed with.) What could have singled out the AGCU alphabet for information transfer is the existence of chemical complementarity relationships between A and U, on one hand, and G and C, on the other. These relationships, with T (thymine) replacing U in DNA, now dominate all forms of information transfer among nucleic acids, as well as the three-dimensional structures of these molecules. In the late 1940s, the Austrian-born American biochemist Erwin Chargaff of Columbia University in New York, spurred by the rising interest in DNA, analyzed DNA samples of widely different origins for their content in their four constituent bases. He made the surprising observation that, within the limits of experimental error, the adenine content was always equal to that of thymine, and the guanine content to that of cytosine.2 It was left for Watson and Crick to discover the fundamental significance of these relationships. In the double helix, the two strands are linked by bonds between A and T, on one hand, and G and C, on the other. The two strands are com¬ plementary because one always has an A where the other has a T, a T where the other has an A, a G where the other has a C, and a C where the other has a G. If one knows the sequence of one strand, one can write the sequence of the other. Watson and Crick also noted that such base pairing could underlie replication. This brilliant intuition has been fully corroborated by later experiments, and extended to RNA, where uracil replaces thymine as the base complementary to ade¬ nine. Thus, in their universal form, the “Chargaff equations” are now written: A === T (or U) and G === C

THE AGE OF INFORMATION

58

These relationships express the ability of the bases concerned to join by their free edges—they are flat molecules—like two pieces of a puzzle, to form flat bimolecular associations stabilized by a special kind of bond known as the hydro¬ gen bond. Hydrogen bonds govern many important interactions between and within biomolecules. The most fundamental of these interactions are those that cause base pairing in nucleic acids. Two hydrogen bonds exist between A and U (or T). There are three between G and C, which thereby make up the stronger of the two pairs. Even so, hydrogen bonds are relatively weak with respect to the forces that tend to destabilize base pairs, namely thermal agitation and electrostatic repulsions between the negatively charged phos¬ phate groups of the two chains. However, the links become cooperatively stronger as the number of adjacent bases engaged in pairing increases. Thus, any two nucleotide chains that contain complementary stretches of, say, three or more nucleotides can be cemented into relatively stable associations by base pairing. These associations require the two chains to face each other by their bases, which implies that the chains are aligned in antiparallel orientation, with the phos¬ phate end of one adjoining the ribose end of the other, and vice versa (see the struc¬ ture on p. 23). Owing to the particular molecular anatomy of polynucleotide chains, the connected stretches are twisted into a helical coil, somewhat in the shape of a spiral staircase. The base pairs form the steps of the staircase, their planes perpen¬ dicular to the main axis and rotated horizontally with respect to each other. The phosphate-ribose backbones of the chains form the banisters of the staircase. When two polynucleotide chains are complementary over their entire lengths, they join into a long, regular, helical, double-stranded thread stabilized by base pair¬ ing. This conformation was first discovered in DNA, which exists in nature mostly in double-stranded form. This is the famous double helix. Unlike Watson and Crick, na¬ ture discovered the double helix first in RNA. In contrast to DNA, most RNA mole¬ cules are single-stranded, except in certain viruses. However, single-stranded RNAs include many short nucleotide sequences that are complementary in antiparallel ori¬ entation. These stretches can join by base pairing, causing the slender RNA chains to form loops of varying complexity, generating structures that have been poetically de¬ scribed as clover leaves, flowers, and the like, but probably appear in reality more like hideously twisted knots. Whatever their aesthetic value, these shapes play im¬ portant roles in determining the functional properties of the molecules. Most likely, the first RNA molecules occasionally included such complementary stretches, which caused the chains to form loops. This may be how replication was initiated.

THE START OF REPLICATION Consider a random RNA chain ending, for example, with the sequence GACU. Occasionally, an A unit in the body of this chain-one out of sixty-four, on aver-

RNA TAKES OVER

59

age—will be followed by GUC. This AGUC sequence is complementary to the ter¬ minal sequence written in antiparallel fashion, and will cause the chain to double up as follows: -C-C-A-G-A-G-U-C-^

: : :

: : : : : :

: : :

\ \ 1

U-C-A-G-'' Assume now that this folded chain is subject to elongation, by the addition of new nucleotides, from right to left, to the U end. The presence of G next to the A paired with the'terminal U is likely to favor the addition of a complementary C over that of the other three possible nucleotides. Repeat the process and you get U added opposite A, G opposite C, G again opposite the next C, and so on. What you get is the formation of a stretch complementary over all its length to the other end of the molecule: -C-C-A-G-A-G-U-C\

:::::::: ::::::::

\

J

-G-G-U-C-U-C-A-G-'' Cut off the loop at the C and G on the right and you get two entirely complemen¬ tary chains. The same would happen if a short GACU stretch served as “primer” for elongation with a “template” molecule ending with AGUC (without a connecting loop). This, essentially, is what happens in biological RNA (and DNA) replication. In this process, a polynucleotide chain is assembled along an antiparallel, complete chain that serves as template. At each step, the nucleotide capable of base-pairing with the one facing it on the template is selected from the four kinds of available nucleotides. The mechanism is extremely simple. It comes down to choosing, from a total number of four, the piece of a puzzle that fits a model. A child of three could do it. Blind molecular agitation until a chance fit locks the correct molecule into place achieves the same result in nature without the child’s observation power. The final outcome is a chain fully complementary with the template. You might say that this is not replication. However, just repeat the process with the newly formed chain as template and you will obtain a copy of the first template. In other words, RNA (and DNA) replication takes place in the same way as photo¬ graphic replication: A positive is made from a negative, and a negative from a posi¬ tive. Double-stranded nucleic acids contain duplex copies of the same molecular (sequence) information, one in positive, the other in negative form. Their replica¬ tion takes place in a reciprocal manner, with the positive serving as template for the assembly of a new negative, and vice versa, to give two identical duplexes as the final result. This is the true significance of the Watson-Crick discovery. It is contained in a single sentence at the end of their 1953 note in Nature: “It has not escaped our

THE AGE OF INFORMATION

60

notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material.”3 The American science writer Horace Judson has called this sentence “one of the most coy statements in the literature of science.”4 Purportedly written by Crick, the English member of the team, the sen¬ tence could also be described as one of the most famous British understatements. Just imagine the inner excitement of the two young men who were no doubt aware that they had just unlocked one of nature’s most jealously guarded secrets. The emergence of RNA was a truly revolutionary innovation in the development of life. By opening the way to molecular copying, it made possible for the first time a mechanism of evolutionary self-improvement by variation, competition, and selection. Henceforth, the historian trying to reconstruct the emergence of life is entitled to look for new explanations besides chemical determinism. A novel phe¬ nomenon had arisen to guide events: Darwinian selection.

DARWIN PLAYS WITH MOLECULES Imagine the setting. The prebiotic soup has “cooked”—via protometabolism—to the point that ATP, GTP, CTP, and UTP (and perhaps other similar molecules) are all being made and combined into a variety of polynucleotide stretches. Among these stretches, some happen to possess appropriate complementary sequences allowing them to double up or join in a manner that permits their replicative elon¬ gation. Such favorable sequences will be replicated and will become more abun¬ dant than the others. For the first time, a mechanism of selection by amplification has arisen, thanks to replication. But this is not all. RNA replication must have been a fumbling affair at first. Inevitably, many mis¬ takes were made, resulting in the formation of many inexact complementary copies of the template sequences. Since these variants could themselves serve as tem¬ plates, a further dispersion of the sequence information ensued. Not each variant, however, gave the same number of copies. Because of a particular feature of their sequence, some were replicated faster than others and were advantaged. Stability under the prevailing conditions was another favorable factor. Molecules combining these two assets—replicatability and stability—in optimal fashion produced the most progeny, which, being endowed with the same advantages, progressively crowded out less advantaged sequences. At the end of this process, a single sequence must have come to dominate the whole mixture, however complex the starting situation. This scenario is not just a theoretical construction. It has been re-enacted many times in the laboratory, first in 1967, by the late American biochemist Sol Spiegelman, from Columbia University, a pioneer in this field.5 What Spiegelman did was to mix in a test tube RNA extracted from a small virus called Q(3, the RNA-replicat-

RNA TAKES OVER

61

ing enzyme (replicase) from the same virus, and the four substrates (ATP, GTP, CTP, and UTP) of RNA replication. After a brief incubation, during which RNA replication took place, he used the RNA formed to seed another similar mixture, repeating the procedure a .number of times. In the end, the resulting RNA turned out to be very different from the viral RNA added in the beginning. It had been stream¬ lined to retain only those parts of the molecule that were needed to ensure effective interaction with the replicating enzyme and to preserve stability. A different final product was obtained when the experiment was performed in the presence of an inhibitor that changed the conditions required for optimal interaction with the enzyme. Since these historic experiments were performed, they have been vari¬ ously repeated many times, especially by Orgel6 and by the German Nobelist Man¬ fred Eigen and his coworkers at the Max Planck Institute for Physical Chemistry in Gottingen, who have also made a detailed theoretical study of the system.7 What is observed is no less than authentic Darwinian evolution at the molecular level. There is a gene, an RNA molecule of given sequence, able to undergo repli¬ cation. This gene is subject to mutations. The mutants compete with each other for available resources—the limited quantity of nucleotides usable for replication. The winners are those that multiply fastest. The important point is that this result is achieved without any design or foresight. The mutations are caused by replication accidents, fortuitous events that bear no relationship to the production of better replicators. This is the essence of Darwin’s theory. Natural selection operates blindly on material offered to it by chance. At the end of the optimization process just described, the system settles into what is known as a steady state, a state of apparent stability in which replication and breakdown compensate each other and in which the optimized sequence main¬ tains its supremacy thanks to continuing selection. Even in the optimized steady state, the RNA molecules are not all identical. Because replication errors continue to occur, the outcome is a continually shifting population of molecules, which Eigen has called a quasispecies. This population consists of perfect copies of the optimal sequence (“master sequence”), accompanied by a covey of variants gener¬ ated by replication errors. It is believed that, in the origin of life, the master sequence of the first RNA quasispecies characterized the first gene. This was an utterly “selfish” gene—to use an expression coined by the British biologist Richard Dawkins8—geared only to its own replication. Eigen has tried to draw the portrait of this primordial gene on the basis of all the information available from experiment and theory, somewhat in the way police artists draw the pictures of wanted criminals on the basis of testimonies.9 His “iden¬ tikit” picture of the first gene shows a surprising similarity to the image that can be deduced from comparative sequencing data for the common ancestor of all extant members of a special class of small RNA molecules, known as transfer RNAs, that play a key role in protein synthesis. Because of the many uncertainties affecting both kinds of reconstruction, no judge would yet pronounce a conviction on the strength of this identification. But it is highly suggestive. The possibility that the

THE AGE OF INFORMATION

62

primordial gene might have been the ancestor of transfer RNAs carries implications of utmost importance and interest with respect to the manner in which RNA mole¬ cules first became involved in peptide assembly. The first outcome of the kind of molecular selection just described may have been the selection of the four bases that make up RNA today. Molecules made ex¬ clusively of A, G, U, and C had the advantage that they could be replicated thanks to base pairing. Molecules containing other bases unable to engage in base pairing were weeded out.

THE BIRTH OF PROTEINS With the appearance of the primordial gene, nascent life, having exhausted the pos¬ sibilities of molecular improvement by Darwinian selection, must have worked itself into what looked like an evolutionary dead end, the only means of escape being a change of external conditions. Even then, freedom would not have lasted long. The selective pressure would have changed, but soon again the system would have run into another optimization blind alley, adapted to the new conditions. This would indeed have been so but for the occurrence of a new kind of reaction that ini¬ tiated further progression. According to the most likely scenario, it all began with the emergence of one or more RNA variants that interacted with amino acids in such a way that the amino acid became linked to the ribose end of the RNA molecule. This is how the primor¬ dial gene started the long evolutionary journey that eventually gave rise to transfer RNAs, its closest extant descendants according to Eigen’s identification. For this is exactly what transfer RNAs do today. They join with amino acids in what happens to be the first step of protein synthesis. Emerging life did not “know” that this interaction was opening the way to one of its most momentous developments: RNA-dependent protein synthesis. There must have been an immediate advantage whereby RNAs capable of interacting with amino acids were selected for preferential replication. The explanation could be simple. RNAs with amino acids hooked to their tails could have folded into a more compact conformation that protected them against breakdown. Or they could have served more efficiently as templates for replication—not an implausible possibility since the amino acids were attached to the end of the RNA molecules where read¬ ing of the template starts. The presence of the amino acid could have ensured that reading started at the proper place, or it could have facilitated the interaction of the template with the catalytic system in some other way. Thus, simple Darwinian selection at the molecular level could have provided the driving force, not only behind RNA evolution, but also behind the involvement of RNA in protein synthe¬ sis, one of the most crucial events in the history of life. Attachment of amino acids to transfer RNAs requires energy, which, in nature, is supplied by ATR In my model, the energy could have come from thioester bonds

RNA TAKES OVER

63

if the amino acids had reacted as thioesters. There are other possibilities, including a role for ATP, already present in the system at the time, and even a direct interac¬ tion beween RNAs and free amino acids.10 An interesting question is whether the interactions were specific. Did a particular kind of RNA join specifically with a particular kind of amino acid? Or did the same RNA react with several different amino acids, or the same amino acid with several different RNAs? In protein synthesis today, the interaction between transfer RNAs is highly specific. However, this specificity is ensured essentially by the enzymes that catalyze the association. These enzymes possess binding sites that specifically rec¬ ognize a given amino acid and a corresponding transfer RNA, positioning the two molecules in such a way that they become linked together with the help of ATP. There is little evidence that transfer RNAs and amino acids recognize each other directly without the help of the joining enzymes, although this possibility cannot be ex¬ cluded: Certain direct RNA-amino acid interactions have been observed.11 It is tempting to assume that the primitive process exhibited some specificity, either by direct interactions or through the mediation of some catalytic surface. This would explain the puzzling selectivity of protein synthesis, which uses only twenty distinct amino acids and leaves out many others that are available, including some that probably were abundantly present in the primeval soup. Also, there is the intriguing fact that nineteen out of the twenty proteinogenic amino acids—the twentieth, glycine, does not exist in two forms—are “left-handed.” Molecular handedness—the technical term is “chirality,” from the Greek cheir, hand—is a property of pairs of molecules that, like our two hands, are constructed identically except that one appears in space as the mirror image of the other. The two forms are designated D and L, the initials of the Latin words for right {dexter) and left {leavus). Proteins contain only L-amino acids. This strange preference of nature for left-handed amino acids is viewed by many scientists as one of the most intriguing mysteries surrounding the origin of life. It is conceivable that primitive transfer RNAs were instrumental, by their specificities, in selecting certain amino acids for protein synthesis.12 Furthermore, it is difficult to explain the origin of translation and of the genetic code without assuming that the primordial couplings between amino acids and RNA molecules enjoyed a certain degree of specificity. Once amino acid-carrying RNA molecules were roving around in increasing abundance, it is to be expected that they started to interact with each other. This is what happens today to amino acid-carrying transfer RNAs. In a first step, two such molecules confront each other in such a way that the amino acid of one joins with the amino acid of the other to form a dipeptide. Then, through a similar type of interaction, the dipeptide gains another amino acid provided by a transfer RNA and becomes a tripeptide. This act is repeated many times, until a particular polypeptide chain is completed. Proteins are assembled by this mechanism in the whole living world. It seems very likely that assembly of peptides was inaugurated by the pri¬ mordial amino acid-bearing RNAs and that RNA-dependent protein synthesis was born in this way.

64

THE AGE OF INFORMATION

In nature, peptide assembly takes place on ribosomes, which are highly com¬ plex, compact particles made of several kinds of RNA molecules (ribosomal RNAs) and more than fifty different proteins. The protein-synthesizing machinery is completed by a thread of messenger RNA, which dictates which of the twenty amino acids is to be inserted at each step. However, this last mechanism need not yet concern us, as the code according to which it operated was not yet developed. We are dealing with an uninstructed form of peptide synthesis. Even if we leave out the information aspect, the prominent role of RNAs in present-day protein synthesis remains striking. Impressed with this fact, Crick suggested in 1968 that the first protein-assembly machinery might have consisted exclusively of RNA molecules, without proteins.13 This was not an unreasonable proposal since proteins could hardly have been available initially for making the machinery that was going to make them. The later discovery of catalytic RNAs gave a great boost to Crick’s suggestion, which has become one of the main props of the RNA-world model. Even though the thioester-world model allows for the intervention of catalytic multimers in primordial peptide assembly, we cannot ignore the eloquent message from nature. It seems very likely that RNAs ancestral to ribosomal RNAs and, perhaps, to messenger RNAs were involved as structural and catalytic components of the primitive peptide-assembly machinery. The find¬ ing by the American investigator Harry Noller, of the University of California at Santa Cruz, that the ribosomal catalyst responsible for sealing peptide bonds may itself be of RNA nature, provides strong additional support to this hypothesis.14 What is not clear, however, is how RNAs catalytically active in peptide assem¬ bly came to be selected. Perhaps involvement in this process somehow favored the replicatability or stability of the molecules. But this explanation is not very con¬ vincing. In any case, there was soon to be a need for a new selection mechanism based on the usefulness of the assembled peptides. This will be the subject of the next chapter.

Chapter 6

The Code

We have reached a point in our hypothetical reconstruction of the age of information where the first peptides began to be assembled by an RNA machinery. We know what came next: translation and the genetic code. Two questions chal¬ lenge the historian. First, by what succession of steps did translation and the genetic code arise? Second, what was the driving force that propelled such an extraordinary development? The two questions are intimately related, since no pathway can be considered that does not entail an explanation of its spontaneous emergence. Before we try to answer these questions, a new element needs to be introduced, namely, the concept of a primeval cell.

DARWIN NEEDS CELLS The cell is the unit of life and figures at some stage in all attempted reconstructions of the origin of life. Some scenarios bring in cells early or, even, right at the start. Others begin with an unstructured soup and introduce cellularization later, some¬ times postponing it to the last moment before it became indispensable for further progress. For reasons that will be explained in chapter 9, I have adopted the latter course. But a limit has been reached. With the initiation of RNA-dependent peptide synthesis, if not before, emerging life had virtually exhausted the potential of molecular evolution. For further evolu¬ tion to take place, less selfish criteria for selection—or, better said, less crudely self¬ ish criteria—had to come into play. RNA molecules no longer had to be assessed solely on the strength of their intrinsic ability to survive and be replicated, but on the basis of their ability to do something that favored their survival and replication indi¬ rectly. But for this kind of selection to operate, the biogenic system needed to be parceled out into a number of discrete, semiautonomous, self-reproducing units—let us call them protocells—each containing its individual genome. Then, any useful

66

THE AGE OF INFORMATION

mutation would benefit only the protocell in which it occurred, causing this protocell to reproduce faster than the others, together with its improved genome, and to pro¬ gressively squeeze out the others by the ever-augmenting throngs of its similarly ad¬ vantaged progeny. In order not to break the thread of my narrative, I shall examine the mechanisms that led to the appearance of the first protocells in a subsequent chapter (chapter 9). I shall assume, for the time being, that cellularization has occurred and that the events we are about to consider took place in a population of protocells capable of individual growth and of reproducing by division. Once protocells existed, selection could proceed on a wider basis and favor any replicatable RNA that, one way or the other, enhanced the capacity of its protocellular proprietor to grow and produce progeny. It is at this stage, and not earlier, that catalytic RNAs could have been retained and improved by selection, to the extent that their activities happened to be useful to the protocells concerned. In particular, there would have been a strong selective pressure in favor of more efficient parts of the peptide-synthesizing machinery if, as appears likely, the ability to make pep¬ tides was an asset in itself. A considerable additional advantage would have been provided by anything that favored the making of useful peptides, as against useless or harmful ones. But for this to happen a feedback loop was needed whereby the useful peptides selectively promoted their own replication.1 Direct copying of the useful peptides could con¬ ceivably have done the job, but there are good reasons for rejecting this possibility. Perhaps the best reason for doing so, unless evidence to the contrary becomes available, is that protein copying does not solve anything. We are still left with the problem of explaining translation.

ANATOMY OF TRANSLATION Our clue, once again, comes straight out of present-day life. The protein-synthesiz¬ ing machinery consists of several parts. First, there is the ribosome, which is the cat¬ alytic assembly bench. It is a small, dense particle, one-millionth of an inch in size, constructed of a large and a small subunit, each made of RNA and protein compo¬ nents in roughly equal proportion by weight. The ribosome automatically joins an amino acid to a growing peptide chain (a single amino acid to start with) when appropriately presented with these two molecules. In terms of information, the ribo¬ some is illiterate. It acts blindly and links any two partners that are in the right chem¬ ical conformation and are properly aligned in regard to its catalytic center. A second part of the machinery is messenger RNA. It runs like a tape between the two ribosome subunits, which it helps to keep together, and it provides the information that specifies the order in which amino acids are to be assembled in the course of peptide synthesis. In this translation from nucleic-acid language to pro-

THE CODE

67

tein language, the nucleotide sequence of the messenger RNA stipulates the aminoacid sequence of the corresponding peptide. This takes place in simple colinear fashion: Each consecutive triplet of bases (codons) in the messenger RNA specifies in the same order an amino acid of the polypeptide chain. Of the sixty-four different codons that can be constructed from the four different bases, sixty-one specify one of the twenty amino acids out of which proteins are constructed, and three are stop codons signifying the end of assembly. A special amino-acid codon doubles as the starting codon. These one-to-one relationships between a triplet of bases and an amino acid make up the genetic code. With minor exceptions, the same code is obeyed throughout the living world. It is a universal dictionary. With its associated messenger RNA, the ribosome turns into an assembly bench that makes a single type of polypeptide. As a rule, ten or more ribosomes, busily reading the message and making the corresponding polypeptide chain, follow each other on the same messenger-RNA strand. Such strings are called polysomes. Tens of thousands of polysomes, assembling thousands of different proteins, are present at any given time in any given cell. Amino acids are conveyed to these machineries by transfer RNAs. Properly aligned on the ribosome surface, two amino acid-carrying transfer RNAs interact so that one donates its amino acid to the amino acid attached to the other transfer RNA, which now bears a dipeptide. Subsequent lengthening of the chain takes place by what the late German-born American biochemist Fritz Lipmann has called “head growth.”2 At each step, the entire growing chain is transferred to the next amino acid brought in by its transfer-RNA carrier (see figure 6.1). Imagine a train being assembled, not by hooking on one car after another to the tail of the train, but by moving the whole lengthening train each time to attach it to the next head car, ending with the locomotive. In this way, the lengthening peptide chain continually remains attached to a transfer RNA by means of its most recently acquired amino acid until synthesis is completed, at which time a stop codon signals the detach¬ ment of the finished polypeptide chain from its final carrier. The sites on the ribosome by which amino acid-carrying transfer RNAs are bound recognize features shared by all transfer RNAs, whereas the catalytic center joins chemical groups common to all amino acids. Discrimination is carried out entirely by the messenger RNA. In performing this function, the messenger RNA does not “see” the amino acids. It sees only the transfer RNAs. More precisely, it sees only a small part of that RNA, a triplet of bases, or anticodon, complementary to the codon. The location of the anticodon in the transfer-RNA molecule is such that it is correctly aligned in antiparallel fashion along the codon exposed by the messenger RNA when the transfer-RNA molecule occupies a binding site on a ribo¬ some. The association between codons and anticodons depends on base pairing, with some leeway, or “wobble,” at the third base of the codon, which allows a given anticodon sometimes to interact with more than one codon. (There are about forty transfer RNAs for the sixty-one amino-acid codons.) At each step of the assembly process, the ribosome and messenger RNA together shape a newly

FIGURE 6.1

The Main Steps in Protein Synthesis

1.

-A-C-

C-G-- (mTNA) (tRNAs)

= catalytic site)

3.

>=i>

1. A transfer RNA (tRNA) bearing amino acid E is aligned on the ribosome next to a trans¬ fer RNA bearing the growing peptide. 2. The growing peptide is transferred from its transfer RNA (which detaches from the ribo¬ some) onto amino acid E presented by the neighbor transfer RNA. 3. The ribosome has shifted along the messenger RNA (mRNA), and the next amino acid (P) is presented by a transfer RNA with a complementary anticodon. Note that the peptide elongates by head growth.

THE CODE

69

opened site, within which only one among the forty-odd transfer RNAs involved in protein synthesis can fit correctly with its attached amino acid. This is how the mes¬ sage is read. It is the puzzle game all over again, but with forty pieces instead of the four used in RNA replication. It may take a child of five to do it successfully. It is a characteristic of this process that the reading is done entirely in RNA lan¬ guage by means of base pairing between codons and anticodons. The translation step proper is carried out before assembly, by the enzymes that attach amino acids to transfer RNAs. These enzymes recognize both an amino acid and a correspond¬ ing transfer RNA. They are the only parts of the translation machinery that under¬ stand both “proteinese” and “RNAese,” albeit only a single word of each language per enzyme. If ct mistake is made by one of these enzymes and it attaches the wrong amino acid to a given transfer RNA, the assembly machinery has no way of detect¬ ing the mistake. It slavishly obeys the instruction provided by the anticodon of that wrongly loaded transfer RNA and adds the wrong amino acid to the growing chain. Astonishingly, only about half of the “bilingual” enzymes that attach the appro¬ priate amino acids to the appropriate transfer RNAs recognize the anticodons on these transfer RNAs.3 The other half of the set of enzymes involved in translation recognize particular transfer-RNA features other than the anticodons and are unaf¬ fected by changes in the anticodon or even by its complete removal. They are thus bilingual by proxy, so to speak—their RNA specificity does not come straight out of the genetic dictionary. It relates instead to structural elements on the transfer RNAs distinct from the anticodon and sometimes separated from it by a consider¬ able distance. This is a very puzzling fact, as it adds another link in the information chain and therefore increases the chance of error. It is difficult to understand why evolution would select such an unnecessary weak link, which seems likely to be a vestige of an earlier relationship that evolution failed to erase. I shall come back to the possible nature of this relationship. But first let us look at the development of translation itself.

THE ORIGIN OF TRANSLATION To understand how a machinery as intricate as that involved in protein synthesis could ever have come into being, let us imagine a situation where no code existed and peptides were assembled in random fashion. You might expect that messenger RNAs, being meaningless, would not be needed in such a situation. This would not be so. Even in the present-day machinery, messenger RNAs, or their equivalent, would still be required because they play a conformational role in addition to their informational role. They help lock the two loaded transfer RNAs together with the two ribosomal subunits in the conformation required to transfer the growing pep¬ tide chain to the next amino acid. This, I submit, explains the entry of messenger RNA on the peptide-synthesizing

70

THE AGE OF INFORMATION

scene. Its precursor was part of the original catalytic RNA scaffolding on which the first peptides were assembled. In this scaffolding, the primitive RNA destined to become ribosomal RNAs provided the catalytic part, and the primitive RNA that evolved into messenger RNAs ensured that the amino acid—carrying and peptide¬ carrying RNAs were properly positioned, accomplishing this by the same kind of triplet interaction that now exists between codons and anticodons. For simplicity’s sake, I shall use the terms “codon” and “anticodon” to designate these triplets, and I shall refer to the three RNA species involved by the names of their present-day descendants: ribosomal, messenger, and transfer RNAs. Together, these three types of RNA made the first peptides, assembling amino acids in an order that, if not entirely random, was far from being as strictly imposed as it is with today’s set of RNAs. Some of the peptides made in this way were use¬ ful. Therefore, the mere ability to make peptides was an asset to the protocells con¬ cerned, and any mutation of the RNAs that enhanced this ability conferred a selec¬ tive advantage on the protocells. Thus, the various RNAs were jointly subject to Darwinian evolution and selection, with the efficiency of peptide synthesis as the screening criterion. The machinery might improve in this way, though not yet its products. There was no way for a good product to feed back positively on its pro¬ duction. The seeds of such a feedback did, however, exist in the fact that codonanticodon interactions were involved in the mechanism whereby messenger RNAs helped immobilize loaded transfer RNAs in the appropriate orientation for peptide synthesis. Messenger RNAs thus, from the start, made a selection among the trans¬ fer RNAs at each step, the degree of specificity of this selection depending on the number of different transfer-RNA molecules that carried the same anticodon. Such being the case, we may expect natural selection to weed out ambiguities and lead to a situation where each kind of amino acid becomes attached to a transfer RNA bearing a specific anticodon. Imagine, for example, a case where two transfer RNAs carrying the amino acids glycine and alanine, respectively, have the same anticodon: GGC, that is, the sequence guanine-guanine-cytosine. Wherever the codon GCC (complementary to GGC in antiparallel orientation) appears in the messenger RNA, either glycine or alanine will be inserted in the forming peptide, with chance deciding between the two. Now let a chance mutation in the transfer RNA for alanine change the central G of the anticodon to C. Only glycine is now inserted in regard to codon GCC, whereas alanine is now called for by codon GGC. The system has gained in speci¬ ficity. Let one or more of the new peptides turn out to be useful and the protocells that possess the mutated transfer RNA will enjoy a selective advantage and produce more progeny possessing these more specific transfer RNAs. Should the same mutation affect the transfer RNA for glycine instead of the transfer RNA for alanine, a similar gain in specificity will be achieved, but the posi¬ tions of the two amino acids in the peptides will be reversed. The protocells con¬ cerned may also be advantaged. The final outcome will depend on which of the two

THE CODE

71

sets of peptides confers the greatest evolutionary advantage. According to the genetic code, the first set was better. GCC happens to be a codon for glycine, and GGC a codon for alanine. This kind of scenario could be repeated with other amino acids and other trans¬ fer RNAs. Eventually, all twenty proteinogenic amino acids would progressively be pulled into the system. Evolutionary step by evolutionary step, translation and the genetic code would emerge concomitantly as products of natural selection. The proposed mechanism requires that each transfer RNA be specific for a particular amino acid. This is consistent with the view that transfer RNAs “fished” out the proteinogenic amino acids. Even if the original specificities had been relatively loose, natural selection would progressively make them stricter. An interesting feature of the proposed model is that natural selection started screening peptides almost as soon as a primitive peptide-synthesizing machinery was set in place. The mutations that played an important role in the beginning were those that affected the anticodons of transfer RNAs and thereby changed the sequences of all the peptides containing the amino acid carried by the mutated transfer RNA. Whole sets of peptides with given positions occupied by either one or another amino acid were subjected to screening by natural selection. Later, as translation and an unambiguous code progressively emerged, mutations of the transfer RNAs became almost invariably lethal because the consequences were too widespread to be tolerable. Among the many peptides that were altered simultane¬ ously, some were almost bound to become defective in the process. The motor of evolution shifted to mutations in messenger RNAs. These mutations led to the syn¬ thesis of only one kind of altered peptide, which could then be evaluated on the basis of its usefulness. Most often, the altered peptide will have been inferior to its nonmutated predecessor and the affected protocells eliminated in the competition. Occasionally, the mutation would bring improvement and give the affected proto¬ cells a selective advantage. This mechanism (with DNA eventually replacing RNA as the mutable storage form of the information) has become the central driving force of evolution. There is a third possibility. The altered peptide was neither better nor worse than its predecessor. The mutation was neutral and carried along passively by what is called genetic drift. These are the mutations that allow us to reconstruct the tree of life by comparative sequencing.

STRUCTURE OF THE CODE A central issue, still unresolved, is whether the structure of the genetic code is a product of chance or was imposed by deterministic factors. Put otherwise, if living organisms similar to those on the Earth existed elsewhere, would they use our code or another one?

72

THE AGE OF INFORMATION

The answer to this question would be straightforward if there happened to exist a direct structural correspondence between amino acids and their anticodons, that is, it the first transfer RNAs had actually fished out amino acids using their anti¬ codons as hooks. The code would then be strictly deterministic. Many efforts have been made to uncover such relationships, almost invariably in vain. Although not entirely hopeless, the prospects of this line of research do not look encouraging. The fact remains that primitive RNAs and amino acids must have “seen” some¬ thing in each other. Why would they have come together otherwise? Furthermore, that something must have been different for each transfer RNA-amino acid combi¬ nation, to account for the specificity without which the emergence of translation seems difficult to explain.4 It is an intriguing possibility, so far unconfirmed, that the transfer-RNA features involved in these early recognitions may be related to the structural traits recognized today by those enzymes that ignore anticodons in the selection of the transfer RNAs to which they bind amino acids. This would explain the evolutionary retention of such features in some cases, and in other cases, their erasure and replacement by anticodons as the recognized features. Be that as it may, the necessary recognition between primitive transfer RNAs and amino acids implies that the starting situation was far from random and could well be the same should the process be reproduced elsewhere. Further evolution through mutations affecting the anticodons would have to operate within those con¬ straints. Perhaps amino acids would end up coded by the anticodons that specify them in present-day life, but this is far from certain. Another important element is the historical factor. It is likely that the twenty proteinogenic amino acids were not all available initially. The code, therefore, must have started with a smaller number of amino acids—estimates vary between four and eight—and must have emerged progressively as more amino acids entered the circuit. Several different models have been proposed to account for this emergence. All the models have in common that they impose limits on the kinds of anticodons likely to represent given amino acids. To take a simple example, the German chemist Manfred Eigen5 has speculated, for reasons that we need not go into, that primitive transfer RNAs might be made of repetitive GXC triplets, with X being any one of the four available bases: G, C, A, or U. There is room for four anticodons in such a struc¬ ture—GGC, GCC, GAC, and GUC—which, in antiparallel orientation, correspond to codons GCC, GGC, GUC, and GAC, respectively. In the present-day world, these codons specify the amino acids alanine, glycine, valine, and aspartic acid, respec¬ tively. These amino acids happen to be the most abundant among the proteinogenic amino acids formed in Miller’s simulation experiments or found in meteorites. It is difficult to see this as mere coincidence. Whether Eigen’s scenario or another is the correct one, the important point is that chance and selection had to work within a severely constrained historical con¬ text. Amino acids were encoded in the order in which they became available for peptide synthesis, whereas the codons themselves were assigned in an order that most likely was not random but imposed by molecular exigencies of the RNAs

THE CODE

73

involved. In other words, codons were distributed among amino acids—or amino acids among codons—on a mutual, “first come, first served” basis. It is impossible to estimate the stringency of these constraints, but their existence makes it very probable that the structure of the code is not purely accidental, as is sometimes claimed. There is another aspect to the code that suggests a nonrandom origin. Its struc¬ ture is remarkably regular. Codons coding for the same amino acid or for amino acids with similar properties are grouped together in such a way that the harmful consequences of mutations (due to chance replacement of one base by another in the triplet) are minimized. In many instances, the altered codon calls for the same amino acid or for an amino acid sufficiently similar to the original one for the prop¬ erties of the altered peptide not to be significantly modified. This regularity sug¬ gests that the code was molded by natural selection during the long period when protocells were experimenting with different codon assignments and vying with each other for leadership in the evolutionary race. In conclusion, it is not certain that aliens would understand our genetic lan¬ guage, but the odds that they would do so are far from negligible. To be true, evolu¬ tion has played a few tricks with the genetic code since the code was first estab¬ lished, for example, in mitochondria, a characteristic part of eukaryotic cells. This, however, was a very late event that happened when less than a dozen genes were left in the mitochondria to be affected by the change. It tells nothing of the histori¬ cal constraints that shaped the code in the course of early evolution.

METABOLISM REPLACES PROTOMETABOLISM The development of translation and the genetic code only opened the way out of the RNA world. There followed a long period during which protocells progressively acquired new peptides. Let us imagine how this happened. Invariably, the first initi¬ ating step was the occurrence of a chance mutation in some RNA molecule. Remember, in the RNA world, RNA molecules served both as replicatable genes and as translatable messenger RNAs. Therefore, the mutation was hereditary and expressed as a new peptide. If this peptide gave the protocell in which the mutation occurred an advantage over the other protocells in the Darwinian “struggle for life,” the protocell and its progeny would multiply faster than the others and progres¬ sively take over. The same series of events must have taken place hundreds of times in succession before a protocell emerged that was fully able to survive and multiply with the help of its newly acquired peptide armamentarium. Only then could this protocell population dispense entirely with whatever supported its ancestors in the RNA world. What properties of the new peptides could have been useful to the protocells so

74

THE AGE OF INFORMATION

as to induce their selection? In the majority of cases, catalytic properties must have been the main assets that singled out peptides for selection. These assets had to be appraised within the framework of existing protometabolism. A catalytic peptide, even of exceptional activity and specificity, that did not find a substrate to act on or that gave rise to an unusable product would have been no good to the protocell in which it arose and would not have been retained by selection. In contrast, a catalyst that fitted within the scheme of things would have been good material for selection, especially if it did a better job than the existing catalyst or if it extended the net¬ work of metabolic pathways into new directions that led to selective advantage. This brings us to the important point emphasized in part I, namely, the need for congruence between early protometabolism and present-day metabolism. The net¬ work of chemical intermediates participating in protometabolism served as a pow¬ erful screen for the selection of the appropriate enzymes among the offered pep¬ tides. Protometabolism could gradually evolve into metabolism, and multimers give place to enzymes, without transgressing the sacrosanct Central Dogma. There was no need for peptide replication or for reverse translation of primitive peptides into the corresponding RNAs. All the required information was present in the meta¬ bolic network. Metabolic superhighways were not constructed independently of existing country roads, but came about by widening and resurfacing those country roads slowly over time. Concurrently with the stepwise development of translation and the genetic code, enzymes manufactured by the RNA machinery, or rather their peptide precursors, progressively took over the jobs previously carried out by primeval catalysts. The transition was gradual, as it took a long time for translation to reach a stage where peptides were made reproducibly from well-defined, replicatable RNA blueprints. It also took a long time for protocells to acquire hundreds of enzymes, one by one, by mutation-selection. Protometabolism progressively gave way to metabolism during that time, but it could give over completely only after the last essential enzyme had been put into place. During this transition, my hypothetical multimers, if they ever existed, became increasingly dispensable. However, the capacity to make multimers from thioesters may not necessarily have disappeared. That capacity could have been retained and perfected by mutation-selection if some multimers happened to carry out useful functions that none of the new peptides could accomplish. The synthesis of grami¬ cidin S and other odd peptides from thioesters by certain bacteria could be a her¬ itage from the ancient ancestral mechanism. It could, however, also be a recent evo¬ lutionary reinvention. Thioesters play so many important roles in all living organisms that their use for peptide synthesis could easily have emerged more than once.

Chapter 7

Genes in the Making

Between

the appearance of the first peptide, haphazardly assembled by

interacting RNA molecules, and the inauguration of a fully integrated translation apparatus, complete with an unambiguous genetic code and a reliable set of func¬ tional RNAs and enzymes for enforcing the code, emerging life went through a long succession of tiny evolutionary jumps separated by more or less extended periods of random groping. An image that comes to mind is that of a surface of water slowly spreading over an irregular terrain. Fingerlets extend here and there, as local attractive forces battle with surface tension, until a minor breakthrough suddenly occurs in a given direction and all the pressure momentarily concentrates on a single rivulet. After that, groping soon resumes, sending out feelers until the next breakthrough. Evolution did its groping by means of chance mutations resulting in the synthe¬ sis of altered peptide molecules; it achieved its directional jumps through the occa¬ sional emergence of an altered peptide product that conferred a selective advantage on the protocell concerned. As with spreading water, the outcome of such a process depends on the structure of the terrain. Without better knowledge of the prebiotic terrain, we cannot reconstruct this phase of evolution in any detail, but we can guess its eventual outcome with some measure of confidence. By the time this phase came to an end, most, if not all, of the twenty proteinogenic amino acids had been recruited for peptide synthesis, the genetic code had reached its present struc¬ ture, except for possible minor adjustments, and translation of RNA messages into peptides was essentially unequivocal and reliable. What, then, were the next steps?

THE MODULAR GAME Most likely, at this stage, genes still consisted of RNA. Those early RNA genes were short, no more than seventy to one hundred nucleotides long (the length of

76

THE AGE OF INFORMATION

present-day transfer RNAs). This estimate follows from the rule, established by Eigen,1 that the number of unit building blocks in a replicatable macromolecule cannot exceed the inverse of the error rate of the replication process. Otherwise, the information content of the molecule becomes irretrievably lost upon repeated repli¬ cation. It is estimated that crude RNA replication had an error rate of one nucleotide wrongly inserted for every seventy to one hundred nucleotides added, depending on the base composition of the RNA molecule. Hence the estimate of seventy to one hundred nucleotides for the maximum length of the first genes. It follows that the peptide products of the first genes cannot have been more than about twenty to thirty amino acids long—one amino acid for every nucleotide triplet—allowing for some noncoding parts in the genes. These peptides were retained by natural selection. Therefore, they had some useful function, most often a catalytic one. This tells us two things. First, peptides that short can display enzyme-like catalytic activities—a point of importance with respect to the multimers of my model. Second, enzymes did start as relatively short peptides. This fact counters the argument, often proffered by creationists, allegedly prov¬ ing that life cannot have originated by a natural process. Consider, it is said, a pro¬ tein, such as cytochrome c, made of one hundred amino acids. Imagine that at each step in the synthesis of this protein the amino acid to be added is decided by the throw of a twenty-faceted die (one facet for each of the twenty proteinogenic amino acids). The chance that the right amino acid will be selected is one in twenty at each step. For the whole sequence of one hundred amino acids, the probability that the correct sequence will turn up is one in 20100, or one in 10130—for all practical pur¬ poses, zero. And cytochrome c is one of the shortest of several thousand proteins present in any cell. Hence the conclusion that life cannot have arisen by a natural process. Fet us, however, repeat the calculation for a peptide of twenty amino acids and let us assume further that only eight different amino-acid species are available for its formation, as might have been the case at an early stage of evolution. Each dis¬ tinct possible sequence then has a probability of one in 820, or one in 1018. It would take only one billion billion protocells, which could comfortably fit within a small pond if they were the size of bacteria, to try out all possible sequences. Even if all twenty proteinogenic amino acids were used, the number of protocells needed for a complete survey, which is on the order of 1026, could still be accommodated in a small lake. In other words, if proteins started as short peptides, emerging life could have explored the totality of what is called the sequence space, leaving nothing to chance. The reader may have detected a flaw in this reasoning. Granted that the number of possibilities stays within manageable limits for the first twenty amino acids, another eighty amino acids remain to be added in order to arrive at cytochrome c. Thus, the original objection still appears to be valid. Cytochrome c cannot have arisen by chance. This would be true if the next eighty amino acids were added one by one. But

GENES IN THE MAKING

77

they were not. The next phase in the evolution of proteins very likely took place by a combinatorial game using existing peptides (by way of their genes) as modular construction blocks. This fact completely changes the probabilistic outlook. Assume, for instance, thqt a set of one thousand peptides of twenty amino acids was selected during the first phase of protein evolution. With these peptides as building blocks, one can construct 1,0002, or one million, different peptides of forty amino acids. All possible combinations are readily tried and submitted to natural selection. Let something like one thousand such peptides eventually emerge and it will again be possible for all combinations of sixty and eighty amino acids to be screened. It thus appears that the whole gamut of present-day proteins could have been created through an exhaustive exploration of the sequence space, provided the expansion of this space by the lengthening of sequences was appropriately pruned by natural selection. The importance of the historical factor in such a process must be underscored. At each stage, evolution can work only with the materials that have survived from earlier trials. Even if a previously rejected combination should turn out to be highly desirable at a later stage, there would be no way of retrieving it, except by chance mutations of existing combinations. The historical dimension of the evolu¬ tionary process is of very general significance. We shall encounter it many times in subsequent chapters. As evolution proceeds in a given direction, the range of avail¬ able choices narrows, and its commitment becomes increasingly focused and irre¬ versible. Considerable evidence of modular construction exists in present-day proteins. The number of modular units out of which all existing proteins have been con¬ structed has even been estimated. The indications, although still subject to consid¬ erable uncertainty, are that this number could be of the order of only a few thou¬ sand. If confirmed, such a figure would be extraordinarily suggestive: the whole variety of life created from permutations of a few thousand building blocks! I shall have more to say on this topic in chapter 24.

R N A SPLICING The modular game was not played by peptides but by the genes that coded for them, most likely RNA genes. A new catalytic armamentarium was needed for this. The most important catalyst required was one that would join, or splice, two sepa¬ rate RNA chains into a single one. In modern jargon, this kind of activity is called trans splicing {trans means “on the other side of” in Latin). By itself, this activity often failed to create coherent messages because the coding parts of the spliced RNAs were not joined in phase—that is, without codon disruption—to allow con¬ tinuous reading, or were separated by a noncoding stretch. In order to correct such defects, a different kind of activity was needed that would cut out a piece between

78

THE AGE OF INFORMATION

two messages and splice them together again, in phase. This kind of splicing within a single molecule is called cis splicing (from the Latin word for “on the same side of”). Finally, proper insertion of the messenger-RNA molecule into the translation machinery may have required some trimming at the end of the molecule (the end of the tape must be cut off for proper fit). All three of these processes occur in many extant organisms, though no longer in the course of RNA gene assembly—RNA genes were phased out long before the appearance of the common ancestor of all present life on Earth—but rather as a means of unscrambling at the RNA level a mysterious phenomenon of gene frag¬ mentation at the DNA level. I shall discuss this phenomenon further in chapter 24. Let it simply be stated that many genes, especially in higher eukaryotes, are split into segments that are expressed, and therefore called exons, and into intervening segments, or introns, that are not. These split DNA sequences are transcribed in their entirety, and the resulting RNA molecules are subsequently processed in such a way that the introns are removed and the exons spliced together. Sometimes com¬ pleted by end trimming, this processing gives rise to mature RNA molecules. These then either become part of some machinery, such as the protein-synthesizing sys¬ tem, or, more frequently, serve as messenger RNAs and are translated into proteins. Split genes are virtually absent in bacteria; they are scarce in lower eukaryotes and more abundant in higher eukaryotes, where their frequency tends to increase with evolutionary advancement. Post-transcriptional RNA processing thus has the appearance of a late evolutionary acquisition. Whether this is so or not—we shall see later that this question is disputed—the processing enzymes could be very ancient heirlooms that go back to the RNA world. It is remarkable that all three of the activities involved—trans splicing, cis splicing, and end trimming—can be cat¬ alyzed by special RNA molecules without the assistance of any protein. Catalytic RNAs, or ribozymes, were discovered by the study of these processes. Proteins are invariably involved as well, but the fact that they are dispensable is viewed as highly significant. It suggests that the relevant activities were originally carried out by ribozymes alone. Thus, next to the translation machinery, a second major cat¬ alytic system arose from interactions among RNA molecules. This fact has supplied another powerful argument in support of the RNA-world model. The interplay among RNA molecules that led to the emergence of RNA splicing presumably went through the usual combination of random mutation and then screening by natural selection. It is likely that protocells equipped with splicing ability gained a selective advantage from some of the longer peptides that were formed through the translation of spliced RNA genes. Replication of the spliced genes must have run into a problem, however, because their length exceeded the limit of seventy to one hundred nucleotides imposed by the error rate of replication. The solution to this problem was the development of more accurate replicating enzymes, which were a prize catch for selection to net. Until this happened, replica¬ tion had to go on using the shorter genes as templates, which implies that protocells continued to rely on splicing to retain the longer, useful peptide products. Thus, any

GENES IN THE MAKING

79

improvement in the specificity, accuracy, and reproducibility of RNA splicing was an advantage. This splicing process still plays a major role today, especially in higher eukaryotes, even though, with the advent of DNA, it is no longer the main mechanism generating variations for the evolutionary combinatorial game.

THE ADVENT OF DNA As their genetic diversity and sophistication increased, protocells must have faced growing logistic problems. Picture the two complementary forms of hundreds of RNA “minigenes” and their splicing products vying for base pairing, replication, splicing, and translation, and you readily visualize the inextricable tangle in which protocells became increasingly snarled the further they progressed. There was only one way out: division of labor. Replication had to be separated from translation. DNA had to emerge. Nobody knows when this crucial development took place, but it seems likely that it happened at a time when the formation of larger RNA genes was already well advanced. Chemically, DNA is a chainlike macromolecule very similar to RNA. It is like¬ wise made of a large number of nucleotides chosen from four distinct species. There are two differences. The sugar ribose is replaced by deoxyribose, which is ribose from which an oxygen atom has been removed—hence the prefix “deoxy” and the name “deoxyribonucleic acid,” DNA for short. The second difference con¬ cerns one of the four bases, uracil, which is replaced in DNA by thymine, which is uracil to which a methyl group (CH3) has been added. This modification does not affect base pairing, so that the pair AT in DNA is equivalent to the pair AU in RNA. The pair GC is the same in both types of molecules. Only minor metabolic innovations were required for the building blocks of DNA assembly, dATP, dGTP, dCTP, and dTTP—d stands for “deoxy”—to become avail¬ able. When these molecules appeared, three key reactions became possible, all ruled by base pairing, as in RNA replication. First came reverse transcription, the assembly of DNA on an RNA template. It is called “reverse” because it was discovered after transcription, the assembly of RNA on a DNA template, which is the main reaction linking these two information¬ carrying molecules in the contemporary world. Historically, however, reverse tran¬ scription most likely emerged first. It played a crucial role by allowing information stored in RNA molecules to be transferred to DNA molecules. Storage without the possibility of retrieval would have been useless. Hence the need for transcription. The stored information could thereby be recovered in a form suitable for translation. This back-and-forth movement of information between a form (DNA) that is unavailable to the translation machinery and one (RNA) that is available to the machinery provided a valuable way of regulating the expression of genetic information.

80

THE AGE OF INFORMATION

Finally, DNA replication, or the assembly of new DNA on an existing DNA tem¬ plate, completed the installation of the new genetic machinery, by entirely dissoci¬ ating the replication of information from its expression. It is likely that these three functions were all performed at first by the same en¬ zyme, which was also the catalyst responsible for RNA replication. The substrates and the templates involved in the four types of reactions were sufficiently similar for a crude catalyst not to discriminate efficiently among them. However, as soon as the use of DNA as a storage form of genetic information provided some selective ad¬ vantage, evolution took over in its usual way, putting mutations to a test and letting natural selection retain whatever happened to be useful. Increasingly, specific en¬ zymes thus arose from mutations of the original gene coding for the multifunctional catalyst of nucleic-acid assembly. Eventually, four distinct enzymes, each specific for a single type of reaction, emerged. They are, in current terminology: RNA replicase, DNA replicase (more commonly called DNA polymerase), transcriptase, and reverse transcriptase. (The suffix “ase” denotes an enzyme.) Once the DNA system was firmly established, the two enzymes using RNA tem¬ plates became useless, even harmful, as they could only confuse matters. It was more advantageous for the protocells to have an unambiguous chain of command, from DNA to RNA to protein, and to restrict replication to DNA. There was thus a considerable evolutionary pressure in favor of eradication of the RNA replicase and reverse transcriptase genes. These have, indeed, largely disappeared from the living world, except in certain viruses. Viruses are infectious agents that can be reproduced only with the help of the chemical machinery of a living cell. Polio, rabies, smallpox, measles are caused by viruses that reproduce in animal or human cells. Viruses that infect plant cells, protists, or bacteria also exist. All viruses have in common a genome carrying their blueprint and the means to introduce this genome into a cell in a manner that allows reproduction of the virus by the cell. Some viruses have DNA genomes like the rest of the living world, but others have RNA genomes. RNA viruses come in two types. In one class (for example, polio), the viral RNA is reproduced by direct replica¬ tion, with the help of an RNA replicase encoded by a viral (RNA) gene. The viral RNA (or its complementary replica) also acts as messenger RNA in the expression of the viral genome. When RNA viruses of the second type infect a cell, the RNA is first subjected to reverse transcription to DNA, with the help of a reverse transcriptase encoded by a viral gene. Transcription of the DNA then serves in expression of the viral genome and in its replication. Such viruses are called retroviruses. They include a number of cancer-causing viruses, as well as the dreaded human immunodeficiency virus (HIV), the causal agent of AIDS, or acquired immunodeficiency syndrome, the plague of the modern world. It has been suggested that viruses are descendants of early forms of life that pre¬ ceded cells. This cannot be so, however, since viruses cannot multiply without

GENES IN THE MAKING

81

cells. They are viewed today as information-carrying remnants or fragments of cells, reduced to the bare minimum required for perpetuation with the help of other cells. Viruses are gypsy genes let loose from their original residences, equipped for wandering from cell to cell, and capable of refreshing and replenish¬ ing their stock at each passage. It is possible that some viruses started their wan¬ dering at a very early stage in the development of life. The RNA viruses, in particu¬ lar, could go back to the days when protocells were getting rid of RNA replicase and reverse transcriptase. The viruses could thereby have saved these enzymes from total eradication. An alternative possibility is that these enzymes were “re¬ invented” at a later stage, for example, by mutation of some DNA replicase or transcriptase gene. Comparative molecular etymology may someday give us the answer to this intriguing question.

GENETIC ORGANIZATION With DNA in charge, many important improvements to genetic organization became possible. First, genes could be stored in single copies or in the minimum number of copies needed to satisfy the requirements of growth. The need for multi¬ ple copies existed particularly for genes coding for structural or functional RNA molecules, such as ribosomal or transfer RNAs. In contrast, messenger RNAs could be generated from single copies of DNA, since translation provided an adequate means of further amplification. As a second advantage, all the genes could now be kept as stable, doublestranded threads, from which one or the other strand, occasionally both, was selected for transcription through the mediation of special, strategically situated nucleotide sequences, called promoters, that control the interaction between the genes and the transcribing systems. As evolution proceeded, these sequences became the target of many regulatory interventions serving to turn transcription of given genes on or off. Transcriptional control of gene expression, which was des¬ tined to become an immensely important mechanism in adaptation and develop¬ ment, was thereby initiated. We shall encounter this mechanism on several occa¬ sions in subsequent chapters. Because genes no longer had to serve as messengers, they could be joined together in strings of increasing length, which, in turn, offered the possibility of a synchronous and appropriately timed replication of all the genes of the string. This development was conditioned by improvements in the accuracy of replication. It is remarkable that much higher fidelity eventually came to be achieved for DNA than for RNA replication. Whereas the lowest error rate for RNA replication is on the order of one in a few tens of thousands—consistent with a maximum length of 20,000 to 30,000 nucleotides for viral RNAs—the error rate can be as low as one in one billion for DNA replication. Due to the existence of elaborate “proofreading”

82

THE AGE OF INFORMATION

mechanisms, whereby wrongly added nucleotides are removed before they are sealed in by the next nucleotide in the growing chain, this remarkable accuracy has allowed all the genes of a bacterial cell, covering millions of nucleotides, to be strung in a single, circular chromosome, which is turned on for replication from a single commanding site, called the origin of replication. It took evolution a great many steps to move from the tangled jumble of small RNA genes to the majestic orderliness of the bacterial chromosome. However, from the moment the first stretch of DNA was assembled, each step provided an incre¬ mental selective advantage. The whole succession followed the characteristic alter¬ nation of mutation and selection that is the modus operandi of evolution.

Chapter 8

Freedoms and Constraints

With the appearance of the first RNA molecules, incipient life entered the

era of molecularly encoded information and progressively built the DNA-RNAprotein triad that now rules the entire biosphere. Three key concepts were intro¬ duced in the wake of informational molecules: complementarity, contingency, and modular assembly.

COMPLEMENTARITY Biological information transfer is based on chemical complementarity, the relation¬ ship that exists between two molecular structures that fit one another closely. Images such as lock and key, mold and statue, are often used to illustrate such a relationship. In the chemical realm, complementarity is a more dynamic phenome¬ non than these images suggest. The two partners are not rigid. When they embrace, they mold themselves to each other to some extent. Furthermore, the embrace leads to binding. Its degree of intimacy is such that electrostatic interactions and other short-range physical forces act strongly enough to prevent the association from being disrupted by thermal jostling. Base pairing, the support of the genetic language, is the most spectacular mani¬ festation of chemical complementarity in biology. But it is only one of many. Every facet of life depends on molecules that “recognize” each other. Self-assembly, the phenomenon whereby complex structures are formed from a number of parts, rests on complementarity relationships between the parts, as did the assembly of furni¬ ture in the old days, except that chemical parts even provide their own glue. Take the immune system and its astonishing versatility and specificity. What makes us resistant against polio or diphtheria—as a result of a previous attack or vaccination—is the presence in our blood of special protein molecules, antibodies, that specifically bind to some component, termed antigen, of the polio virus or of the diphtheria bacillus. The cells that recognize a grafted heart or kidney as foreign.

84

THE AGE OF INFORMATION

and reject it, do so through the mediation of surface molecules that join with some surface component peculiar to the intruder. The white blood cells that stalk invad¬ ing microbes and gobble them up recognize their prey by a similar mechanism. Hormones, drugs, poisons, and every other chemical that exerts a biological effect owe this property to their ability to interact with a receptor molecule on their target. This kind of relationship is now exploited on a vast scale in research. Endor¬ phins, which are natural inducers of pleasurable sensations, were discovered through the morphine receptor. Enzymes offer another fundamentally important example of complementarity. Most enzyme-catalyzed reactions take place in three interconnected steps. First, the molecule or molecules on which the enzyme is to act—its substrate or substrates— become physically bound to special binding sites on the enzyme surface. This bind¬ ing is such that the molecules are offered in appropriate spatial orientation to the catalytic site of the enzyme. Catalysis is the second step, followed, in a third step, by detachment of the products, so that the cycle can start again. As a simple anal¬ ogy, imagine a welder immobilizing two pieces of metal in a vise, then proceeding with the welding, and finally removing the welded product to start a new operation. Alternatively, a single piece of metal could be similarly immobilized prior to saw¬ ing or filing. In enzymatic reactions, there is no workman to select materials. The process is self-powered and depends on molecular affinities between binding sites and sub¬ strates. Thanks to these affinities, enzymes can “fish out” their substrates from highly complex mixtures. In any living cell, hundreds, if not thousands, of different substances coexist, all at very low concentration, as they might in a prebiotic mix¬ ture. Enzyme specificities, as defined by the affinities of the substrate-binding sites of the enzymes, determine the chemical pathways the molecules follow. This relationship can work both ways. Just as receptors may fish out hormones, substrates can select their enzymes, either directly by protective binding—many enzymes are more resistant to degradation when linked to their substrates—or indi¬ rectly through the enzymes’ activities. This is what I believe happened when the RNA machinery started delivering peptides. The catalytic peptides that fitted within the protometabolic network were retained. In this sense, protometabolism already contained information. It provided the blueprint for metabolism through the mecha¬ nism of enzyme selection.

CONTINGENCY Contingency entered the history of life on Earth with the onset of replication and the inevitable mutations that perturb this process. Thus was set in motion the process of Darwinian evolution that has governed the history of life on Earth. Genetic information is accidentally altered. The modified message is replicated and

FREEDOMS AND CONSTRAINTS

85

expressed. The ability of the modified phenotype to perpetuate the modified geno¬ type by means of offspring is evaluated by natural selection, which weeds out harmful mutations with poor reproductive success, favors useful mutations that enhance survival and reproduction, and lets neutral mutations simply drift along. As soon as replication started, this process was initiated, first at the molecular level, then at the level of the protocell. Because mutations are accidental, no two RNA worlds, even in one billion or more, can have exactly the same microscopic history. But what about their macro¬ scopic history? No two streams follow exactly the same course down a mountain, but all may end up in the same valley. We have no Way of answering this question with certainty, but the likelihood is that, in a large number of cases, incipient RNA worlds will issue into an RNAprotein world similar to ours. My main reason for stating this lies in the stringency of the selection factors that came into play at each stage. Behind these factors lurks a great deal of chemical determinism, often based on complementarity. If my proposed reconstructions are correct, the four RNA bases were selected among a number of related products on the strength of their pairing ability, which allowed amplification of the corresponding RNAs by replication. Molecular selec¬ tion, based on optimal replicatability-stability, next led to a reproducible RNA mas¬ ter sequence, as in the Spiegelman-Eigen type of experiment. Chemical interactions between RNA and amino-acid molecules, once again determined by chemical com¬ plementarity, then selected the proteinogenic amino acids and the corresponding transfer RNAs. All very reproducible and leaving little to chance. Also important was the historic factor, which severely channeled the emergence of translation and of the genetic code, itself shaped further by the condition of low¬ est harm to the organism by mutations. Finally, the existing protometabolic network acted as a unifying screen for the first enzymes produced by the machinery. The order in which the enzymes appeared might have varied with the vagaries of muta¬ tions, but the end result would, in each case, be a metabolism largely copied from protometabolism. Later evolution may also have been subject to more deterministic factors than is often surmised, in spite of the increasing role of contingency. Quite possibly, when DNA originated from RNA, driven by the advantages of a separate storage form of genetic information, there were not too many different possible modifications of the RNA molecule that could have ensured specificity, while at the same time sav¬ ing information transfer between the two molecules.

MODULAR ASSEMBLY A third lesson we learn from our reconstructions is the importance of modular assembly. This is a recurring theme in the history of life. Evolution works with pre-

86

THE AGE OF INFORMATION

existing modules—RNA minigenes at the stage we have considered—which are modified and combined in different ways into larger assemblages that are then screened by natural selection. Implicit in this mechanism is the possibility of exten¬ sively exploring the available sequence space at each step, thus further reducing the role of chance. In conclusion, a number of RNA worlds might abort—and perhaps did on our planet—because chance did not provide a necessary mutation. But those that mature would probably lead to a form of life supported by the same basic metabolic processes and ruled by the same DNA-RNA-protein triad and, perhaps, the same genetic code that characterize our own form of life. Furthermore, because of the limited size of the sequence space available to incipient life, the success score is likely to be high.

PART III

THE AGE OF THE .PROTOCELL

Chapter 9

Encapsulating Life

For

a

fully

operational genetic system to develop, emerging life had to

become partitioned into a population of protocells capable of multiplying by divi¬ sion, so that protocells, not simply molecules, henceforth were subjected to natural selection. So far, we have been content to assume that this partitioning took place. Let us now turn back in time and look into the mechanisms of cellularization and into the new properties that the confinement of life within boundaries both allowed and required.

THE TIMING OF CELLULARIZATION There are two conflicting views of the time at which the first cellular structures appeared. A number of scientists, impressed with the fact that microscopic aggre¬ gates or vesicles of various kinds, crudely reminiscent of living cells, can be observed to form under relatively simple conditions, believe that the formation of primitive cells was the seeding event in the origin of life. Extensive laboratory investigations—by Alexander Oparin in Soviet Russia,1 Alphonse Herrera in Mex¬ ico,2 and Sidney Fox in the United States,3 to mention only the most prominent— have been devoted to such artificial “cells,” though without disclosing any plausi¬ ble pathway for the progressive “vitalization” of the structures. Other scientists have defended early cellularization on the grounds that a membranous structure was required for the initial trapping of sunlight energy.4 Yet others find unaccept¬ able for theoretical reasons the possibility that life could have originated in an unstructured “soup.”5 The opposite view is also defended by many. It has been pointed out that the “primeval soup” need not have filled the whole of oceans. Coastal areas, lagoons, ponds, even puddles, could have provided appropriate sites for the soup to thicken

90

THE AGE OF THE PROTOCELL

and evolve chemically. The hindrances an enveloping structure might have posed to the free circulation of biogenic substances are also mentioned in objections to cellfirst theories. Eigen, for example, believes for this reason that “organization into cells was surely postponed as long as possible.”6 The thioester-based metabolic model I have proposed is not readily compatible with early cellularization and fits better with the concept of an initially unstructured soup. It is suggestive in this respect that some metabolic systems generally consid¬ ered most ancient, for example, the system involved in the fermentation of sugar to alcohol, which uses a thioester-linked energy-retrieval mechanism, are situated in the cytosol, or cell sap, the unstructured part of the cell. Thus, the primeval soup, energized by thioesters, may be seen as progressively developing into a sort of extended protocytosol. In the early stages, the need for free exchanges would have given the unparti¬ tioned protocytosol a clear advantage over separate entities subject to the con¬ straints of a peripheral boundary. For encapsulation to take over, the advantages of confinement must have outweighed the drawbacks. This implies that the isolated systems both enjoyed enough autonomy to survive on relatively simple exchanges with their environment and derived a clear benefit from being enveloped. This situ¬ ation was reached, at the latest, when the RNA machinery for peptide synthesis began to come together, since further evolution of this machinery made the exis¬ tence of a large number of competing protocells mandatory. How did the first cell boundaries form, and from what materials? In order to answer these questions, let us again look for clues in present-day organisms.

CELL BOUNDARIES All living cells are surrounded by an unbroken, filmy envelope, called the plasma membrane. Many cells are also partitioned by internal membranes. The universal fabric of biological membranes is the lipid bilayer, a tenuous, double molecular leaflet, about one five-millionth of an inch thick, usually made largely of phospho¬ lipids. Qualified as amphiphilic, or amphipathic (“with two loves”), these mole¬ cules are characteristically composed of two parts with opposite affinities: a hydrophilic (water-loving) head and a hydrophobic (water-hating) tail, also called lipophilic (fat-loving). Hydrophilicity depends on the attractions that exist between electric charges of opposite sign. The water molecule has no net electric charge, but it is electrically polarized. It has a negative pole situated on the side of the oxygen atom, which tends to appropriate more than its share of electrons, and a dual positive pole made up by the two hydrogen atoms, which protrude asymmetrically as partly naked pro¬ tons on the same side of the molecule. As a consequence of this structure, water molecules bind by either their positive or negative pole to any oppositely charged

ENCAPSULATING LIFE

91

or polarized molecules or chemical groups. Water molecules also bind to each other for this reason. Were this not so, water would be liquid only at very low tempera¬ tures; the Earth would be dry, lifeless, and forever barren. Hydrocarbons, the main constituents of petroleum, and all other substances made entirely or mostly of carbon and hydrogen, being uncharged and nonpolar, are hydrophobic. Many such substances exist in the living world. They are grouped un¬ der the term of lipids, which comes from the Greek word for fat. Hydrophobic mol¬ ecules do not really hate or repel water; they are excluded by it owing to the strong tendency of water molecules to join by electrostatic attractions. In the presence of water, hydrophobic molecules are thus driven together by crowding water molecules. The formation Pf such allocations is itself facilitated by hydrophobic-hydrophobic interactions mediated by short-range forces, weaker than electrostatic forces and known as van der Waals forces, from the name of the Dutch chemist who discovered them.7 Each thus keeps to itself. Oil and water don’t mix. The heads of phospholipid molecules owe their hydrophilic character to a nega¬ tively charged phosphate group, often associated with other charged or polar groups. Two long hydrocarbon chains make up the hydrophobic tails of the mole¬ cules. In the presence of water, phospholipids satisfy their two contradictory loves by forming bilayers. In these structures, each of the two layers consists of closely packed molecules, lined up perpendicularly with respect to the plane of the layer (like the bristles of a brush) and oriented in such a way that the hydrophilic heads all face one side and the hydrophobic tails the other. Each layer is thus a very thin sheet one molecule thick, with a hydrophilic face and a hydrophobic face. In bi¬ layers, the two sheets are sandwiched by their hydrophobic faces held together by van der Waals forces, whereas the hydrophilic faces are directed outward in contact with water. Such bilayers thus interpose an oily film between two watery milieus. Phospholipid bilayers are very fluid and flexible. They form a sort of twodimensional liquid, within which the constituent molecules easily slide along each other within the plane of the bilayer. Because of this property, bilayers can mold themselves around any kind of surface and readily adapt to changes in the confor¬ mation of the surface, as often happens with cells. Phospholipid bilayers are always continuous and self-sealing, and therefore always form closed sacs. In this respect, they resemble soap bubbles, with which they share a number of physical properties. In particular, they can join (fusion) or be split (fission) without loss of continuity. Two phospholipid vesicles may fuse into a single one, like two soap bubbles that bump into each other. Conversely, a single vesicle may divide into two, as some¬ times happens to a soap bubble caught in an air drag. A last important property of phospholipid bilayers is their ease of formation. No more than vigorous mechanical agitation, by means of ultrasound, for example, is needed to turn a mixture of phospholipids and water into a suspension of small vesicular bilayers. A whole industry has been built around this phenomenon. Artifi¬ cial phospholipid vesicles, called liposomes, have found many applications as earn¬ ers for cosmetics, drugs, vaccines, genes, and other agents.

92

THE AGE OF THE PROTOCELL

Phospholipid bilayers are impermeable to most water-soluble (hydrophilic) mol¬ ecules. This property makes bilayers excellent boundaries that allow cells to main¬ tain an internal composition different from that of the surrounding medium. But cells cannot survive sealed off from the outside. They must be able to take up nutri¬ ents, get rid of waste products, and respond to environmental signals. These func¬ tions are carried out by proteins inserted into the bilayers. The sequences of membrane proteins are characterized by one or more trans¬ membrane stretches of about twenty to thirty largely hydrophobic amino acids, typ¬ ically coiled into a helical rod called an a-helix. These rods pass through the bilayer, in close contact with the hydrophobic parts, with which they establish links stabilized by van der Waals forces, and serve to position the proteins within the membrane. The other parts of the protein molecules protrude on the outer and inner faces of the membrane. Most cells in both the prokaryotic and eukaryotic worlds are surrounded by peripheral structures external to the plasma membrane, ranging from a fluffy down to massive, rigid walls. These structures serve to support and defend the cell. They act as molecular filters and may fence off an intermediary space, called the periplasmic space, between the cells proper and their environment. A variety of substances, including proteins, lipids, complex carbohydrates, and special con¬ stituents of unique chemical composition, participate in the building of these outer structures.

MECHANISMS OF CELLULARIZATION Phospholipids are complex molecules that could hardly have been available in the primeval soup. But they could have arisen through the development of protometabolism and been piesent in the soup at the time encapsulation became advantageous. It would then have needed no more than some violent storm for vesicular bilayers to form spontaneously in such a soup, the way artificial liposomes arise today in phospholipid-water mixtures exposed to ultrasonic vibrations. Primitive cells could have been born in this way, but only to die almost immediately of starvation because their phospholipid envelopes would not have let through even the simplest of nutrients. It is conceivable, however, that the empty ghosts of stillborn cells provided anchoring points for some metabolic systems and offered a harbor for hydrophobic peptides. Progressive curving of this structure could give rise to a double-membra¬ nous cup, which could further close into a double-membranous pouch once the structure had acquired the necessary systems of transmembrane communication. According to this model, which has been proposed by the German-bom American cell biologist Gunter Blobel, of the Rockefeller University in New York, the first

ENCAPSULATING LIFE

93

cells would have been bounded by a double membrane.8 This happens to be a characteristic feature of gram-negative bacteria (so called because they react negatively to a test devised by a Danish bacteriologist named Gram). It has, indeed, been suggested that gram-negative bacteria may have preceded gram¬ positive organisms, which have a single membrane. The British biologist Thomas Cavalier-Smith, who champions this idea, has adopted Blobel’s model for this reason.9 However, the outer membrane of gram-negative bacteria is very different in structure from the inner membrane, which represents the true cell boundary, or plasma membrane. A possible alternative is that the first boundary was not made of phospholipids but of peptides and other multimers of largely hydrophobic character, which could have formed a looser and more permeable network than lipid bilayers. This is a plausible possibility, as hydrophobic multimers must have been abundant right from the start, considering the nature of many of the available building blocks. Phospholipids could have come later to plug the holes in the boundary and expand it into a more flexible and versatile membrane, as needed communications became established. Whatever their nature, the mechanisms that led to the encapsulation of the first protocells must have been intimately associated with the creation of appropriate passageways allowing the necessary molecular traffic between the protocells and their environment to take place. There are unfortunately no clues to the long suc¬ cession of molecular events that determined this progressive tightening of barriers around increasingly sophisticated means of crossing them. We can only look at the finished product and speculate about its origin. Let us first consider construction.

THE ASSEMBLY OF MEMBRANES Membranes grow by accretion, that is, by the addition of components to a pre-exist¬ ing membrane.10 Thus, de novo synthesis of a membrane needed to occur only once in the history of life, and all subsequent membranes could have arisen from this ancestral membrane by expansion followed by fission. We don’t know whether things happened this way, but it is an intriguing possibility. At least, membranes develop in this manner in the living world today. Once the first membranes arose, any innovation that facilitated the insertion of new components into them was advantageous. For lipids, the simplest and most effective innovation was to have them synthesized right in the membranes, which provided an excellent milieu for housing the hydrophobic building blocks used. Thus, a number of enzyme systems involved in the synthesis and assembly of lipids, especially phospholipids, became associated with membranes. Today, CMP, the cytosine-containing constituent of RNA molecules, is heavily implicated in these processes as a carrier of several key building blocks. If historically signifi-

94

THE AGE OF THE PROTOCELL

cant, this fact suggests that phospholipid membranes came with or after the RNA world, in agreement with the hypothesis of late cellularization. In the case of proteins, adaptations were of a more subtle kind, as the ribosomes on which protein assembly took place were situated in the soluble compartment of the protocells. Homing of proteins to membranes was achieved by means of certain amino-acid sequences, called signal or targeting sequences, typically present in membrane proteins. These sequences were specifically recognized (bound) by cer¬ tain membrane components that served as docking areas for the proteins carrying the right address tag. Consequent to this binding—another typical example of com¬ plementarity—the proteins carrying the tag became inserted into the fabric of the membranes. Two main variations on this theme developed. In one, the targeting sequence occupies the initiating end of the nascent polypeptide chain and joins with the membrane as soon as this end emerges from the ribosome. Called cotranslational because it takes place while translation is still going on, this transfer is revealed by the observation of ribosomes closely apposed to the inner face of bacte¬ rial cell membranes. The second, posttranslational, kind of protein transfer occurs after completion of the polypeptide chain and depends on targeting sequences that may be situated anywhere in the chain.

THE CONSTRUCTION OF OUTER DEFENSES The construction machineries considered so far played an important role in the functional enrichment of the first membranes but contributed little to their struc¬ tural strength. Phospholipid bilayers, even reinforced by proteins, are flimsy fab¬ rics. They are easily torn or damaged by physical or chemical agents and offer vir¬ tually no resistance to osmotic swelling, a phenomenon induced by the inflow of water that occurs when cells are exposed to a medium in which dissolved sub¬ stances are less concentrated than they are inside the cell. This fragility of their sur¬ face boundary severely curtailed the ability of the protocells to withstand outside aggressions and to adapt to different environments. Then, an event happened that exerted an enormous influence on the prospects of life on Earth. Protocells “learned” to build rigid extracellular structures from carbohydrate building blocks. This historical event probably started with the appearance of mechanisms for joining sugar molecules together into chains, or saccharides, of various lengths that served mainly as reserve substances. What we call sugar in everyday language is actually a disaccharide made of two elementary sugar molecules, glucose and fruc¬ tose. Starch is a polysaccharide made entirely of glucose. One readily sees how the ability to store energy-rich foodstuffs as large molecules that could not escape

ENCAPSULATING LIFE

95

through the surrounding boundary provided the protocells with enough advantages to favor selection. It is interesting, and possibly suggestive, that the main carriers involved in saccharide synthesis today are derivatives of UMP, or occasionally of AMP or GMP, that is, typical RNA constituents. Thus, together with phospholipids, polysaccharides could also be products of the RNA world or of the post-RNA world. The next decisive step was initiated by the formation of a new kind of sugar car¬ rier, derived from a substance called dolichol, anchored in the membrane by a long hydrophobic tail. Sugars or saccharide chains were transferred from their nucleotide carriers to the membrane-bound carrier and thereby made to stick closely to the inner face of the membrane. By an intriguing flipping phenomenon, these bulky, highly hydrophilic masses came to be translocated across the hydrophobic barrier of the phospholipid bilayer and to pop up on the outer face of the membrane. There they could be handed over to protein molecules or to other acceptors. In this way, the surface of the protocells became progressively bolstered and defended by car¬ bohydrate parts, which greatly augmented the survival potential of the protocells concerned. The building of outer defenses involved a remarkable molecule that is still pres¬ ent in a large part of the bacterial world today and has all the hallmarks of a surviv¬ ing fossil. Called murein, this molecule consists of sugar molecules and of short heterogeneous peptides that could, according to their structure and content in both D- and L-amino acids, have come straight out of the primeval multimer mixture. These parts are interlocked into a single, huge, meshlike molecule that entirely envelops the cell within a sort of organic coat of mail. Called the cell wall, this structure is remarkably strong and resilient while being sufficiently porous not to impede molecular passage. Murein is broken down by lysozyme, an enzyme that plays an important role in the defense of organisms against invading bacteria. The naked cells, or protoplasts, that are stripped of their wall by lysozyme usually burst osmotically unless the medium composition is such as to prevent the influx of water. On the other hand, the miracle drug penicillin owes its unique therapeutic virtues to its ability to block the building of murein and thereby prevent the growth and multiplication of sensi¬ tive bacteria. As it happens, lysozyme and penicillin were both discovered by the same scientist, the Scottish microbiologist Alexander Fleming, at a time when noth¬ ing was known about the chemistry and synthesis of the bacterial cell wall.11 The wall was further strengthened by the thickening of the murein layer or by the coating of this layer with a membranous skin constructed from special lipopolysaccharide molecules and rendered permeable to small molecules, but not to proteins, by inserted, tunnel-shaped protein molecules called porins. As men¬ tioned, there is a possibility that this second membrane, which characterizes gram¬ negative bacteria, may be a legacy of an early encapsulation stage in which proto¬ cells were enveloped by a double membrane.

96

THE AGE OF THE PROTOCELL

THE NECESSARY INLETS AND OUTLETS The first, inescapable condition of survival in confinement was the possibility for the protocells to take in food from the outside and get rid of waste material. The simplest way in which fully enveloped protocells could fulfill this condition was by means ot pores, mere holes kept open in lipid bilayers by some kind of inserted protein framework. The porins, just mentioned, are an example of such proteins. Next came transport facilitators, which are transmembrane proteins that act as molecular turnstiles for certain specific substances. Like simple turnstiles, facilita¬ tors are passive systems. They open in either direction and give in to the side from which the pressure is greatest. That is, they let substances flow in the direction leading from a higher to a lower concentration. But they do this with a certain degree of chemical discrimination. Many cells, for example, contain a transport facilitator that provides specific passage for glucose molecules. A more sophisticated kind of molecular turnstile is the gated channel, analogous to some of our controlled admittance devices. Gated channels, like facilitators, merely let certain substances of given chemical specificity move through passively, but they are unidirectional and regulated by a gate that needs to be unlocked by some chemical or electrical signal. The next improvement in the building of molecular transport systems was active transport, hooked to a source of energy, usually the splitting of ATP, so that the spontaneous direction of flow could be reversed and substances could be forced uphill, from a lower to a higher concentration. For the protocells involved, such acquisitions meant that they could now fish out rare but essential substances from their surroundings or, alternatively, rid themselves of toxic refuse even in a highly polluted environment. Although there was an energy bill to pay, the gain in survival potential was high enough to tilt the direction of natural selection. Among the substances that could be actively transported into or out of proto¬ cells, a number were ions, that is, electrically charged entities. In many cases, the displacement of ions in one direction is linked to an equivalent displacement of ions of the opposite charge in the same direction, or of ions of the same charge in the opposite direction. The membrane boundary remains electrically neutral. Some¬ times there is no such compensation and the forced transport of ions creates an imbalance of electric charges, or membrane potential, between the two sides sepaated by the membrane. Such pumps—a name often used for ion-transporting sys¬ tems—are termed electrogenic. 1

A paiticularly important electrogenic pump uses the energy supplied by the splitting of ATP to drive sodium ions (positively charged) out of cells and replace them partly (two against three) by potassium ions (also positively charged), with the consequent building of a membrane potential positive to the outside. In eukary¬ otic cells, this potential has come to play a role of exceptional importance, as the

ENCAPSULATING LIFE

97

basis of all bioelectric phenomena, including the functioning of the nervous system in animals. The origin of the sodium-potassium pump is obscure. It is of possible signifi¬ cance that the principal pgsitively charged ion of sea water—and also of animal blood—is sodium. It is conceivable that fenced-off life had to defend itself very early against excessive sodium. Interestingly, some of the most ancient bacteria, known as halophiles, are particularly effective in coping with external sodium. They thereby manage to survive—in fact, to thrive—in concentrated brine. Another kind of electrogenic pump of central importance forces protons across membranes. Proton pumps powered by the hydrolysis of ATP serve in a number of instances to raise the proton concentration, that is, the acidity, of certain intracellu¬ lar or extracellular regions (think of the acid produced in the human stomach). By far the most important function of proton pumps is in energy transfer, which I shall examine in the next chapter.

CELL DIVISION Whatever the mechanism of encapsulation, it had to include the possibility of turn¬ ing growth into division. Without such a link, advantageous mutations could not have turned into selective assets; they would even have been self-defeating. This point is easy to understand. Consider a spherical cell. As it grows, its mass and, therefore, its maintenance and repair needs increase as a function of the third power of its radius. On the other hand, the surface area it has available for the import of nutrients increases only as a function of the second power of the radius. Growth of such a cell must necessarily reach a point where import just suffices for mainte¬ nance and repair. Further growth is impossible, unless the cell becomes asymmetri¬ cal, forms a bud, for example. Burgeoning of the bud would eventually lead to its falling off as a free entity, especially if a self-sealing membrane surrounded the whole system. Thus, any surface property that favored asymmetric growth and bud¬ ding of the first protocells would, if hereditarily transmissible, automatically have been selected. With the development of outer structures, division became a more complex process, dependent on a progressive constriction of the wall into a deepening circu¬ lar furrow. Little is known of the mechanisms controlling this phenomenon, but a link exists between membrane and wall. Bacteria unable to build a wall as a result of exposure to penicillin and protected against rupture grow bigger but do not divide. For division to be of any use, each daughter protocell had to include all that was needed for autonomous survival and proliferation, in particular a full set of genes. At first, this condition was probably satisfied on a statistical basis by the random mixing of the protocell components within their membranous envelope. When the

98

THE AGE OF THE PROTOCELL

genetic material became centralized into a single, circular chromosome, a relation¬ ship was established between DNA replication and division, such that each daugh¬ ter inherited a chromosome. This partition was facilitated by anchoring of the chro¬ mosome to the plasma membrane. Upon initiation of DNA replication, the complex of enzymes and ancillary factors needed for this process was assembled around the anchoring point, and the chromosome was gradually reeled in through this com¬ plex, exiting in duplicated form. After the two resulting chromosomes became dis¬ entangled, each ended up anchored to a different site of the membrane. The furrow initiating division then formed between the two sites, thus ensuring that each daugh¬ ter cell inherited one of the duplicated chromosomes.

Chapter 10

Turning Membranes into Machines >»

Encapsulation was a slow, progressive process, punctuated by many evolu¬

tionary acquisitions. By necessity, the earliest of these acquisitions concerned mostly means of ensuring vital exchanges with the environment. Soon, however, the scope widened. Once phospholipid bilayers were formed, this new fabric turned out to be much more than a convenient boundary. It presented burgeoning life with numerous opportunities for useful innovation. A whole new class of proteins emerged, fitted with one or more hydrophobic sequences that allowed insertion within membranes. Thus immobilized, the proteins could participate in a variety of novel functions that were sufficiently advantageous to favor the evolutionary selec¬ tion of the mutant protocells that made the proteins. By far the most important development of this sort was the putting together of a machinery coupling downhill electron transfer reversibly to proton extrusion. Emergence of this machinery was a truly revolutionary advance in the ability of life to derive energy from environmen¬ tal sources.

PROTONMOTIVE ELECTRON TRANSFER Imagine the following scenario. It may not have happened as depicted, but the sce¬ nario is plausible and tells in a simple fashion how emerging life may have hit upon the invention that completely transformed its future—made this future possible, in fact. Owing to some mutational event, a protocell acquires an electron-carrying mole¬ cule constructed so as to fit within the fabric of the protocell’s membrane. What makes this carrier useful, and favors its selection, is that it can serve as a bridge for electrons across the membrane, between an internal donor and an outside acceptor

100

THE AGE OF THE PROTOCELL

to which the membrane is impermeable. Access to this acceptor is, thus, the protocell’s gain from the mutation. The lunch is not free, however. There is a price to pay: The carrier transports elec¬ trons in the form of hydrogen atoms. This means that if the transaction between in¬ ternal donor and external acceptor involves naked electrons, protons are necessarily translocated together with the electrons. The carrier must pick up protons from inside the protocell

one proton for each electron—to make the electrons transportable as

hydrogen atoms, and it must discharge the same number of protons outside when it delivers electrons to the external acceptor. Thus, electron transfer is obligatorily coupled to proton translocation, and vice versa. One cannot take place without the other. What this amounts to is a reversible, electron-driven proton pump. Foi such a pump to be of any use, the membrane needs to be impermeable to protons. Otherwise, the translocated protons immediately diffuse back inside. If proton leaks are plugged, the coupling between electron transfer and proton trans¬ location becomes an energy link. As electrons are transferred across the membrane, the accompanying protons create a rising imbalance, or proton potential, which, depending on circumstances, is manifested in the form of an excess of external over internal proton concentration; of a membrane potential, positive outside; or of a combination of both. Whatever its physical form, the proton potential tends to oppose the further translocation of protons. The higher the potential, the stronger the opposing force. The coupled process grinds to a halt when the amount of work required to push more protons against the existing proton potential becomes equal to the amount of energy released by the electron transfer. This amount of energy is itself a function of the difference between the energy levels at which the donor gives out the transferred electrons and the acceptor takes them up. Could such a liability be turned into an advantage? Yes, in several ways. Sur\ ival in an acidic medium is an attractive possibility. By definition, acids are hydro¬ gen-containing substances that, when dissolved in water, tend to release free pro¬ tons (the rest of the molecule being left as a negatively charged ion). The higher the proton concentration created in this way, the stronger the acidity (and the lower the pH, an expression swimming-pool owners will understand). From the tangy sour¬ ness of lemon juice (citric acid) or vinegar (acetic acid) to the metal-biting caustic¬ ity of nitric acid, it is all a matter of proton concentration. As discussed earlier, there are reasons for suspecting that life started in or near an acidic environment. Some of the most ancient microbial species belong to the group of thermoacidophiles, which inhabit a very hot and acidic milieu. We saw in chapter 3 that such a milieu could have been conducive to the formation of the first thioesters. Also, it would have released inorganic phosphate (or pyrophosphate) from its insoluble combinations and allowed this essential ingredient of many bio¬ molecules to enter primitive metabolism. There is a difficulty, however, with early life actually developing in such an environment because a number of metabolic intermediates, including several critical phosphate compounds, are extremely sensi¬ tive to hot acid. The existence of volcanic springs and the recent discovery of deep-

TURNING MEMBRANES INTO MACHINES

101

sea hydrothermal vents suggest a possible way out of this quandary. Thioesters could have arisen, and phosphate dissolved, in hot, acidic subterranean waters and come to the surface with pressurized jets that transferred them to milder conditions. Proto¬ cells developing at the edge of such a source could have invaded increasingly acidic waters by acquiring the means to drive out protons and so to keep an appropriately mild internal milieu against the pressure of a strong external proton potential that would take advantage of any weak spot in the membrane to push protons in. The coupling between downhill electron transfer and proton extrusion could be put to use in an alternative manner if the outside proton potential were strong enough to reverse the flow of electrons across the membrane, that is, to force the electrons to move in the uphill direction, from the reduced form of the outside acceptor, turned into donor, to the oxidized form of the inside donor, turned into acceptor. For this to happen, the protocells would need a proton “sink,” that is, a metabolic system capable of consuming the protons that enter through the pump turning in reverse. The same evolutionary advantages would be associated with the acquisition of an ATP-driven proton pump (see the preceding chapter). Whichever pump came first, it would have conferred a substantial benefit to protocells occupying an acidic medium. It is tempting, though admittedly speculative, to take the emergence of proton pumps as another clue pointing to an acidic cradle for life. A dramatic change, no longer linked to outside acidity, occurred when the two kinds of proton pumps—one electron-driven and the other ATP-driven—turned up together in the same protocell membrane. Imagine the scene. The two pumps start by acting in concert, joining efforts to build a rapidly rising proton potential. Because the two pumps are not likely to be of exactly equal strength, a stage will be reached where the weaker one stops, while the stronger one goes on driving out protons, raising the proton potential above the weaker pump’s limits. When this happens, the weaker pump starts running in reverse. The proton potential built with one source of energy mediates the replenishment of the other source of energy. If the electron-driven proton pump is the stronger of the two, downhill electron trans¬ fer supports the assembly of ATP from ADP and R. If the ATP-driven proton pump is the stronger, ATP hydrolysis supports uphill electron transfer, from a lower to a higher energy level. A new form of reversible coupling between electron transfer and ATP assembly, based on protonmotive force, is born. The importance of this event can hardly be overestimated. Before it happened, the reassembly of ATP (or pyrophosphate) with the help of electron-transfer energy took place entirely by the thioester-dependent mechanism of substrate-level phos¬ phorylation (see chapter 3). Today, probably less than one molecule of ATP in one million is reassembled by this mechanism (which nevertheless remains universal and vitally important). The membrane-bound mechanism of carrier-level phosphor¬ ylation now dominates biological energy retrieval. Without it, we could not cover our energy needs by the combustion of foodstuffs. Nor would plants be able to har¬ ness the energy of the sun.

102

THE AGE OF THE PROTOCELL

The evolutionary advantage of the new energy-retrieval mechanism was imme¬ diate. As soon as its most primitive seed was planted, every improvement in the efficiency and versatility of protonmotive coupling was strongly favored by natural selection. The crowning achievement of this long evolutionary development is the electron-transfer chain—also called the respiratory chain, because, in all aerobic organisms, the electrons are collected at the end of the line by molecular oxygen, which is itself made available by respiration. Such a chain consists of a number of electron carriers, arranged within the fabric of a membrane in a manner that has been compared to an electron bucket brigade or to an electron cascade. The bucketbrigade image underscores the participation of carriers in the flow of electrons along the chain. The image of a cascade makes clear that the pathway followed by the electrons is downhill and includes steps where the electrons fall down a sub¬ stantial difference in energy level. One or more of these steps—three in all most advanced systems

are obligatorily coupled with the extrusion of protons and can

serve to power the assembly of ATP by way of protonmotive force. Thus, when electrons fall down the cascade, ATP molecules are assembled. Conversely, elec¬ trons can be forced up the cascade with the expenditure of ATP or with the help of protonmotive force generated by electrons flowing down the lower part of the chain. A number ot important membrane-embedded molecules participated in the con¬ struction of electron-transfer chains. We have already encountered in chapter 3 the gioup of iron-sulfur proteins, built around iron-sulfur clusters, which operate by way of ferrous/ferric oscillations. Also dependent on the same oscillations are a number of membrane-bound, red-colored substances called cytochromes, which are members of the larger group of hemoproteins, of which the prototype is the red blood pigment, hemoglobin (from the Greek haima, blood). The active part of hemoproteins is a complex, flat, dish-shaped organic molecule, made of carbon, nitrogen, and hydrogen atoms, belonging to the porphyrin group. In the center of the dish is a hole occupied by an iron atom. In cytochromes, this iron atom alter¬ nates between the ferrous and the ferric form, and thus accounts for the electroncarrier function of the molecule. Cytochromes are found in membranes as members of electron-transfer chains. The hemoproteins also include a number of soluble sub¬ stances, in which the iron remains permanently in the ferrous or ferric form. The ferrous hemoproteins mostly act as oxygen carriers, like blood hemoglobin. The ferric ones exert some enzymatic activity involving hydrogen peroxide. In addition to members of these two groups of iron proteins, electron-transfer chains also include cuproproteins, with copper as the electron carrier; flavoproteins, with FMN or FAD (see chapter 4) as the electron carrier; and electron-carrying quinones, highly hydrophobic organic molecules composed of carbon, hydrogen, and oxygen atoms. Altogether, as many as fifteen distinct carriers may be associ¬ ated in a given chain, physically organized in decreasing order of energy level, so that each canier is strategically positioned with respect to the carriers with which it transacts direct electron transfers.

TURNING MEMBRANES INTO MACHINES

103

THE ATTAINMENT OF AUTONOMY It is commonly believed that early life drew its building blocks from preformed organic products of abiotic syntheses. As to its energy needs, they could also have been covered by preformed energy-rich molecules, such as inorganic pyrophos¬ phate or polyphosphates, or, as in my model, thioesters. Alternatively, the break¬ down of preformed organic molecules could have supplied the necessary energy through some coupled process, such as the thioester-generating electron-transfer process envisaged in my model. If such was the case, life started in heterotrophic form. The term “heterotroph” (Greek heteros, other; trophe, food) designates organisms that, like ourselves, feed on products made by other organisms, by contrast with autotrophs (Greek autos, self)—plants, for example—which manufacture their constituents from mineral building blocks. Early heterotrophy did not rely on autotrophs, of course, but on the celestial manna of abiotic chemistry. By the time the manna became exhausted, some form of autotrophy had to be developed. We don’t know when this happened, but it is not likely to have been before protocells appeared, unless some unknown kind of mechanism was involved. All known autotrophs depend on membrane-embedded electron-transfer chains. It is likely, therefore, that such chains started by supporting heterotrophic processes and became converted to autotrophy later. How did this conversion take place? To answer this question, we must look at the anatomy of the biological electron cascade. It consists of four distinct “chutes,” separated by five energy levels—say, in the order of decreasing altitude: A, B, C, D, and E. Each chute is high enough to support the assembly of ATP from ADP and Pj? at the rate of one molecule of ATP per pair of electrons falling down. The A-B chute is the oldest; it involves watersoluble components and depends on thioester-linked, substrate-level phosphoryla¬ tion. The B-C, C-D, and D-E chutes involve membrane-embedded components and depend on protonmotive, carrier-level phosphorylation. For the cascade to operate, electrons must be fed into it and collected from it. Feeding occurs optimally at level A and collection at level E, but electron inlets and outlets also exist at intermediate levels. This cascade is not a pre-existing feature of nature. It is the product of evolution, which, guided by natural selection, succeeded in putting the A-E span to optimal use, within the limits imposed by the energy requirement of ATP assembly. In brief, when two electrons fall from level A to level E, enough energy is released to power the assembly of four ATP molecules altogether. The cascade, with its four chutes, each harnessed to a separate machinery, exploits this possibility to the full. Our problem is to explain how this masterpiece of natural selection came into being and how it paved the way to autotrophy. There is a simple answer to this question. It may not be the correct one, but it

104

THE AGE OF THE PROTOCELL

will do for our purpose. In terms of the image of an electron cascade, it can be sum¬ marized as follows. Life started at the top of the cascade and harnessed first the fall ot electrons from level A to level B or below. This harnessing took place by way of thioesters and evolved into the mechanism of substrate-level phosphorylation. Abi¬ otic syntheses, perhaps helped by a source of high-grade electrons (see chapter 3, The Case of the Missing Hydrogen”), supplied the appropriate electron donors, some ot which still fulfill this function 3.8 billion years later. For example, pyruvic acid, which readily arises from lactic acid or alanine—two characteristic abiotic products—is one of the major electron donors in substrate-level phosphorylations today. As for the collection of electrons, it seems safe to assume that the prebiotic world offered a choice of mineral electron acceptors operating at level B or below. Even organic molecules, produced by either abiotic syntheses or protometabolism, could have done the job, as we know from present-day metabolism.1 According to my scenario, emerging life was supported in this way up to the stage of genetically independent protocells capable of making proteins and compet¬ ing with each other on the strength of their protein innovations. A turning point was reached with the appearance of the first membrane-embedded, two-pump machin¬ ery capable of using the energy released by falling electrons for ATP assembly by way of protonmotive force. Most likely, this machinery received electrons at level B, which is the level at which the majority of metabolic electron donors feed electrons into electron-trans¬ fer chains today, mostly by way of NAD (see chapter 4). Entry at level C is not excluded, since a few metabolic intermediates—succinic acid, for example, a typi¬ cal product of abiotic syntheses—deliver electrons at this level. Oxygen, the uni¬ versal final electron acceptor of present-day life, was not available in the prebiotic world to collect the electrons at the bottom of the chute, but other acceptors could have been piesent, for example, ferric iron (see chapter 3), which, like oxygen, accepts electrons at level E. As a means of satisfying energy requirements, the new machinery was far supe¬ rior to the pre-existing thioester-dependent machinery, since level-B electron donors are much more numerous than level-A donors. But this is only a trivial advantage compared to another, truly life-saving consequence of the new develop¬ ment: The thioester-dependent machinery could now be reversed with the help of ATP provided by the protonmotive machinery. Electrons could be lifted from level B to level A—which is the key level for biosynthetic reductions. The assimilation of carbon dioxide, for example, requires electrons delivered at level A. The term ‘reverse electron transfer” designates the energy-dependent lifting of electrons from a lower to a higher energy level. Once initiated, this kind of bootstrapping

of electrons could continue. A sec¬

ond protonmotive chute, fed at level C and unloading electrons at level D or below, could have appeared, thereby opening the possibility of lifting electrons from level C to level A—first from level C to level B by reversal of the first protonmotive machinery, and then from level B to level A by reversal of the thioester-linked

TURNING MEMBRANES INTO MACEIINES

105

machinery. In all such cases, a lower part of the cascade provides power for lifting electrons to the upper part of the cascade. This, I submit, is how autotrophy arose and emerging life was freed of its depen¬ dence on abiotic chemistry. All that was still needed for authentic autotrophy was for the environment to provide a suitable mineral electron donor delivering elec¬ trons somewhere between levels B and C. Then, in the presence of an appropriate acceptor at level D or below (oxygen today, perhaps ferric Ton in prebiotic times), downhill electron transfer in the lower part of the cascade could cover all the ATP needs, whereas uphill (reverse) electron transfer in the upper part of the cascade, powered by ATP assembled in the lower part, could provide the high-grade elec¬ trons required for the reduction of mineral building blocks, such as carbon dioxide, nitrogen or nitrate, sulfate, and so on. This mode of life characterizes a number of autotrophic bacteria, called chemoautotrophs, which use mostly sulfur derivatives such as hydrogen sulfide or elementary sulfur—likely components of the prebiotic environment—as electron donors. Once the basic skeleton of autotrophy was in place, it is easy to see how further improvements could have arisen, up to the formation of a complete, three-chute, protonmotive, B-C-D-E electron cascade, such as is found today in all the most advanced organisms, be they autotrophic or heterotrophic. The ultimate improvement in electron-transfer chains occurred when a por¬ phyrin variant arose in which a magnesium atom came to replace the iron atom in the central hole of the dish. Thus, presumably, was born chlorophyll (from the Greek chloros, green, and phyllon, leaf), the green pigment of photoautotrophic, or more simply, phototrophic, organisms (Greek phos, light). The immediate benefit of chlorophyll sprang from its ability to garner energy from sunlight and to undergo what is known as an electron delocalization in the process. That is, one of the electrons in the chlorophyll molecule is displaced by the absorbed light energy from its resting level to a level of higher energy. The mole¬ cule is said to be excited by light. This phenomenon occurs with many colored sub¬ stances (which are colored for the very reason that they absorb part of the light they receive), but is most often short-lived. The displaced electron promptly falls back to its resting level and the absorbed energy is given out as heat—not a useless com¬ modity, as we know, but one that cannot be used effectively for the performance of biological work. In the case of the excited chlorophyll, this dissipation of the light energy is avoided. Thanks to an intimate association of the molecule with a membraneembedded respiratory chain—it was born from a member of such a chain, remem¬ ber—the delocalized electron is diverted and led to fall productively through this chain. The light energy thus supports the generation of protonmotive force and, through this force, the assembly of ATP. When they are used for ATP assembly, the electrons boosted by light energy eventually return to the chlorophyll molecule at their resting level, ready for another light-driven lift up. The electrons can thus cycle endlessly—up with light,

106

THE AGE OF THE PROTOCELL

down the cascade—and support ATP assembly in the process (cyclic photophos¬ phorylation). In addition to this cyclic process, a noncyclic process also exists whereby the boosted electrons are used for the many biosynthetic reductions required tor autotrophic life. In this event, chlorophyll recovers the diverted elec¬ trons from some outside source. The evolution of phototrophy went through two main stages. First to appear was photosystem I, which picked up electrons from mineral donors, such as certain sul¬ fur compounds, and boosted these electrons, with or without the additional help of thioester-dependent reverse electron transfer, to the energy level (A) required for biosynthetic reductions. Then came photosystem II, which has the ability to remove electrons from water molecules, with the release of molecular oxygen. Photosystem II lifts the electrons up to an intermediate level of energy from which they can be taken up further by photosystem I and pumped to the top. This development freed autotrophic life from its dependence on an external electron donor. Henceforth, life needed only light and water to turn air and a few water-borne minerals into a luxu¬ riant green mantle, which could, in turn, support many heterotrophic forms of life. This momentous event did, however, also signal the appearance of molecular oxy¬ gen on our planet, with long-term effects of dramatic magnitude (see “The Great Oxygen Crisis,” pp. 135-36).

Chapter 11

Adaptation to Life in Confinement

With cellularization, for the first time life became a property of discrete,

autonomous, individual units capable of diversification. Darwinian competition was the most immediate consequence of this development, as well as the main dri¬ ving force of its further evolution. In addition, cellularization permitted a number of acquisitions that further enhanced the capacity of the protocells to survive and proliferate as individual units, and, therefore, offered likely catches for natural selection to net. In most cases, these adaptations were the result of modifications of the cell membrane, which, in addition to its functions as boundary, site of con¬ trolled passage, and protonmotive machinery, evolved into a sensitive interface capable of exchanging many signals with the environment and initiating appropri¬ ate responses. A brief review, not necessarily chronological, of these aids to protocellular life follows.

SENSING Protocells probably first “learned” to explore their environment chemically, by “tasting” it, so to speak. Here is how it may have happened. It is known that mem¬ brane proteins often have one end sticking out on one side of the membrane, and the other end on the other side, with the hydrophobic transmembrane segment in between. Imagine now that the outer part of such a protein possesses a site comple¬ mentary to a given substance and, therefore, has the ability to bind this substance. Imagine, further, that this binding induces a conformational change—a coiling, for example, or an uncoiling—of the transmembrane segment, such that the inner part of the protein, in turn, undergoes a change in shape. Imagine, finally, that this alter¬ ation of the inner part of the protein causes a specific effect, for example, the open¬ ing or closing of a channel, or the activation or inhibition of an enzyme. What you are witnessing, with the appearance of such a protein, is the creation of a link that

108

THE AGE OF THE PROTOCELL

allows an outside chemical to influence internal events without entering the proto¬ cell, clearly a valuable acquisition if the response turns out to be adaptive—which decision will be left for natural selection to make. One can think of such a protein as a switch controlled by a chemical trigger. The triggering substance is called an agonist, or active agent. The outside part of the switch that binds the agonist is termed a receptor. The internal responding part is the effector. The switch is off (effector inactive) when the receptor is vacant, and on (ef¬ fector activated) when the receptor is occupied by the agonist. There are usually many switches of any given kind on the surface of a single cell. This allows for a graded response between all switches off and all on. The extent of the response de¬ pends on how many of the total receptor sites are occupied by agonist molecules. This proportion depends on the abundance of agonist molecules in the environment and on the avidity with which receptor sites fish them out. With a properly tuned sys¬ tem, a fine adaptation of the response to the amount of agonist present may obtain. As a simple example, imagine a protocell equipped with a channel allowing the entry of a substance that is useful or even essential in small amounts, but becomes harmful in excess. Such a protocell would lead a precarious existence, strictly dependent on finding the right concentration of substance in the environment. Let the protocell now acquire a receptor that binds the substance and is connected to an effector that closes the channel when the receptor is occupied. As the substance’s concentration in the outside medium increases, the number of receptor sites occu¬ pied, and therefore that of channels closed, increases. Fewer channels are open, but each lets in more molecules of substance per unit of time, so that the total number of molecules of substance entering per unit of time remains approximately the same over a wide range of concentration. For the protocells concerned, acquisition of such a protein means the ability to adapt to considerable fluctuations of the sub¬ stance in the environment, a powerful selective advantage. The tiansmembrane receptor-effector combination has enjoyed an enormous evolutionary success, linking an immense variety of substances to a great diversity of responses, including the seeking and catching of food, the avoidance of noxious substances, the triggering of cell division, the stimulation of secretion, and many others. The outcome is especially intriguing when the agonist is produced by another cell. The receptor then makes it possible for the agonist-producing cell to influence events in the receptor-bearing cell. What takes place is communication between cells by means of chemical signals, a phenomenon that has played increas¬ ingly important roles in later evolution, especially in eukaryotes. As you read this sentence, billions of cells communicate with each other through chemical signals in your brain to make this array of printed symbols intelligible. There is no way of knowing when protocells first acquired transmembrane pro¬ teins with useful receptor-effector combinations, or what these first transmembrane proteins were. It is likely, however, that such acquisitions were made almost as soon as the degree of structural and functional sophistication of the protocells allowed, since the resulting selective advantages would have been considerable.

ADAPTATION TO LIFE IN CONFINEMENT

109

MOTILITY Transmembrane proteins c^n also serve to transmit chemical signals from the inte¬ rior of a cell to its surface. One particularly intriguing such molecule has an energy¬ consuming inner part connected to a mobile outer part in such a way that the energy spent inside is converted into mechanical work outside, like an arm connected to an oar or an engine to a propeller. The source of energy could be the splitting of ATP or, alternatively, protonmotive force. In the bacterial world, the ultimate machine of this sort, named flagellum (plural, flagella; Latin for whip), is a rotating helical rod that emerges on'the surface of the cell and is connected to an internal “turbine” driven by protonmotive force. The shaft of this rod traverses the cell membrane and the cell wall through special, tight-fitting “bearings.” No doubt, this elaborate engine was preceded by simpler motors, possibly constructed with ATP-splitting proteins that bend when they bind ATP, and straighten when the bound ATP mole¬ cule is split. Eukaryotic motor systems, including our own muscles, are all built with proteins endowed with this ability. Properly arranged on the cell surface, such proteins could cause the cell to move with respect to the surrounding water. This sort of motility started as a random walk. Cells moved for a while in a given direction and then “tumbled,” resuming movement in another direction. There was little advantage to such a property, which was as likely to bring the cell to a less favorable as to a more favorable envi¬ ronment. Things changed when the motor systems became coupled with chemical receptors. The coupling was primitive, the variable affected being merely the fre¬ quency of tumbling. Receptors sensitive to useful substances came to inhibit tum¬ bling, so that the cells continued to move toward the substance for a longer time. In contrast, receptors sensitive to harmful substances precipitated tumbling, thus shortening the time of progress of the cells in the wrong direction. This mecha¬ nism—known as positive and negative chemotaxis, that is, the ability to seek useful substances and to avoid harmful ones—has remained to the present day, at least in the bacterial world. Even though it acted simply by modulating random events and was only marginally effective for single cells, this mechanism was very powerful at the population level. Its acquisition entailed a considerable evolutionary gain.

PROTEIN EXPORT AND THE BIRTH OF DIGESTION In their wanderings, protocells left traces of their passage in the form of discharged waste products and other chemicals. Such traces could conceivably alert protocells to stay away from or congregate around each other. A more specific mode of dis¬ charge arose through a modification of the mechanism whereby nascent proteins

110

THE AGE OF THE PROTOCELL

are targeted to membranes (see chapter 9). Through the combinatorial vagaries of gene assembly, targeting sequences came to be added also to proteins other than those carrying the appropriate hydrophobic sequences needed for nestling within lipid bilayers. A number of soluble proteins were fitted with the right tag and became embroiled cotranslationally or posttranslationally with membranes. It took only minor changes in the machineries involved for insertion to make way for com¬ plete translocation. The outcome was protein export, or secretion. Among the proteins that came to be discharged out of protocells in this way, a particularly favored group consisted of enzymes that split with the help of water the chemical bonds whereby building blocks are linked in natural macromolecules. These hydrolytic enzymes, also known as hydrolases, fragment proteins into amino acids; nucleic acids into nucleotides; nucleotides into sugars, bases, and phosphate molecules; saccharides into individual sugars; phospholipids into their constituents; and so on. Hydrolases play havoc with vitally important substances, and their acquisition must have carried major risks. The protocells in which such enzymes emerged in fully active form were promptly eliminated until chance provided the harmful proteins with a tag that caused them to be exported as soon as they were made. Then the evolutionary disadvantage suddenly turned into an advantage. Enzymes released into the environment by living protocells could break down organic debris left around by dead protocells, and the resulting small molecules could serve to support the nutritional requirements of the secreting protocells. Digestion was born and, with it, the possibility for a living organism to exploit the synthetic activity of another; in other words, the possibility of heterotrophy at the expense of autotrophy. This event had far-reaching consequences. It freed the organisms concerned from the heavy burden of making their own building blocks and gave them greater scope for innovation. Notably, it has been a key step in the emergence of the whole animal world, including the human species. We live, directly or indirectly, on the products of plant photosynthesis, which we digest in our stomach and intestine with the help of extracellularly secreted enzymes. Protocells surrounded by a simple murein wall permeable to protein mole¬ cules

such protocells would correspond to present-day gram-positive bacteria—

had to reside within a stagnant and confined environment in order to benefit from the secretion of digestive enzymes. Otherwise, the enzymes would have been washed away before they could act. In contrast, protocells equipped with a second membrane

corresponding to gram-negative bacteria—kept the secreted enzymes

within the periplasmic space intercalated between the two membranes. There, the enzymes were able to act on such molecules as were let through from the outside by the porins of the outer membrane. Whether membrane-bounded or not, the space surrounding primitive heterotrophic cells may be viewed as the first digestive pocket in the history of life. The first stomach, so to speak, except that it was not situated inside the organism, but the other way round. We shall see later that internalization of this stomach may

ADAPTATION TO LIFE IN CONFINEMENT

111

have played a decisive role in the conversion of an ancestral prokaryote into the first eukaryote (see chapter 16).

A TOUCH OF SEX One last surface acquisition must be mentioned. It consists of long, slender fila¬ ments, up to several times the length of the cells themselves, that adorn the surface of many bacterial cells. Appropriately called pili (singular, pilus; Latin for hair), these structures' serve mainly as anchors allowing the cells to attach to some sup¬ port. They are neither motile nor connected to effectors. But they have some chemi¬ cal specificity and therefore can act as receptors. For example, pili endowed with the right specificity could immobilize wandering cells in the neighborhood of a rich food supply or cause cells to congregate in colonies. It is easily seen how selection would favor the emergence of pili with useful specificities. Certain pili specific for cellular surface components allow cells to touch each other in a special way. To call such contacts between bacteria fondling would no doubt be too anthropomorphic. The fact remains that the development of such pili did lead to sex and thereby unleashed one of the most, if not the most, powerful forces of diversification to drive evolution. Cells possessing sex pili—called male—use such filaments as some sort of molecular penis to copulate with female cells, that is, cells devoid of sex pili. In the course of such conjugation, as it is termed, male cells introduce into female cells a small, satellite circular piece of DNA, called a plasmid, on which, among others, the genes coding for the sex pilus proteins are situated. Not infrequently, a duplicate of a lesser or greater part of the male chromosome is injected into the female cell together with the plasmid. Subse¬ quent recombination of the injected DNA with the recipient cell’s DNA then brings about the formation of hybrid chromosomes made of genes contributed by the two parents. Until this development occurred, the mutations that provided natural selection with materials from which to choose were mostly base replacements due to replica¬ tion errors or to chemical injuries, as well as insertions, deletions, inversions, and duplications of certain DNA (or RNA) stretches. Thanks to conjugation and recom¬ bination, a whole new gamut of hereditary variations involving entire genes or clus¬ ters of genes was offered to selection. The evolutionary game became immensely richer and more innovative. It is amusing that, in this earliest form of sex, males did their job but females were the real beneficiaries of the evolutionary advances. If you are tempted to derive some sort of sexual moral from this fact, note that the first thing males do upon conjugation is to give the females genes (coding for pilus proteins) that turn them into males.

Chapter 12

The Ancestor of All Life

The

evidence that all known living organisms are descended from a single

common ancestor is overwhelming. We cannot exclude the possibility that unknown or poorly known organisms of different origin exist in some remote envi¬ ronment that has remained isolated for a very long time. However, no discovery suggestive of a major break with “our” way of life has yet been made. Until proven otherwise, the hypothesis of single ancestry holds true. In this chapter, I shall try to reconstruct the profile of our common ancestor and to retrace its emergence historically, paying special attention to the shape of its hid¬ den roots. Did the universal ancestor arise as a single shoot along a highly deter¬ ministic pathway that left little to chance? Or was it one of many branches, a branch that simply happened to spread faster, smothered all others, and ended up filling the world with its progeny?

RECONSTRUCTING THE DISTANT PAST The universal ancestor is defined as the organism that existed just before the tree of life divided into two separate branches that have spread out extensions to the pres¬ ent day. This definition distinguishes the universal ancestor from the more primi¬ tive ancestral forms that came before; it also leaves open the possibility of earlier branchings that have left no extant progeny. In principle, drawing a portrait of this ancestral organism is simple: just put together all the properties that are shared by all living organisms. In practice, three caveats complicate matters. First, we must subtract from our picture such shared properties as could have been acquired sepa¬ rately in individual branches after the first forking of the tree of life. Second, we must also subtract properties that appeared solely in one branch and were later acquired by the other branch or branches by some mechanism of gene transfer.

THE ANCESTOR OF ALL LIFE

113

Finally, we must add to the picture the properties that are lacking in certain organ¬ isms, perhaps in all, because they were lost in the course of evolution. The first caveat refers to what is known as evolutionary convergence. This is an important phenomenon in later evolution, illustrated, for example, by the develop¬ ment of flight independently in insects, pterosaurs, birds, and bats, but it is proba¬ bly of minor relevance in the kind of molecular evolution we are mostly concerned with here. It is very unlikely, for instance, that a molecule such as cytochrome c, which shares more than fifty out of one hundred or so amino acids in all species investigated, could have arisen independently in two or more branches. The second caveat is more serious. Horizontal gene transfer1—so named in >*

opposition to “vertical” transfer from generation to generation—is believed to be a common phenomenon in the bacterial world. It seems likely that primitive organ¬ isms exchanged genes at least as readily as do present-day bacteria. However, to account for the occurrence of the same gene in all extant organisms, horizontal transfer of the gene must have occurred very early in evolution, when very few branches (most likely only two) existed in the same or closely connected niches. In addition, members of the transformed branch that did not acquire the gene must have been eradicated. As to the third caveat, it is obviously relevant and must be applied with careful discrimination. Fortunately, enough universally shared features are left to make the main picture fairly clear. Uncertainties concern a few additional properties that are not found in all forms of life but could conceivably have been present in the com¬ mon ancestor and been lost subsequently in some of its descendants. Even with the above caveats, such a wealth of biochemical information has become available on all major forms of life that one would expect the reconstruc¬ tion of the common ancestor to be fairly straightforward. This would be so but for what has come to be known as the rooting problem. In the late 1970s, the American microbiologist Carl Woese,2 from the University of Illinois, dropped what amounted to two simultaneous bombshells on the scien¬ tific world. First, he announced, on the basis of the comparative sequencing of RNA molecules found in the ribosomes of all living beings, that extant bacteria are not members of a single family, as was generally assumed, but fall within two groups that must have separated at the dawn of cellular life. He elevated these two groups to the rank of kingdom and named one archaebacteria, because of a number of characters he believed to be particularly archaic (from the Greek arkhaios, ancient), and the other, eubacteria (from the Greek eu, good). The two kingdoms are grouped together under the name prokaryotes (Greek karyon, kernel), which indicates that neither possesses a true nucleus, in contrast to the eukaryotes, which encompass all protists, plants, fungi, and animals. Woese3 has more recently pro¬ moted the two kingdoms to an even higher rank, for which he has proposed the term “domain,” and has renamed them archaea and bacteria to emphasize their dif¬ ferences. This proposal, however, is not yet generally accepted. I have not followed it in this book because the familiar term bacterium has become so much a part of

114

THE AGE OF THE PROTOCELL

everyday vocabulary that redefining it is likely to be confusing to most readers. On the other hand, I have adopted Woese’s original classification, which met with some resistance at first but is now almost unanimously accepted. Woese’s second bombshell was even more stunning. Eukaryotes, which were commonly believed to have detached from the (single) prokaryotic trunk some time around one billion years ago or later, are almost three billion years older. They orig¬ inated from a line that branched from the tree of life virtually at the same time that archaebacteria and eubacteria separated. Thus, the common ancestor lies at the root of a trifurcation. However, bifurca¬ tions, not trifurcations, trace the development of an evolutionary tree. There are thus three possibilities: (1) The first bifurcation separated archaebacteria and eubacteria, and eukaryotes then branched off the archaebacterial line. (2) Prokary¬ otes separated first, as above, but eukaryotes branched off the eubacterial line. (3) The first fork separated eukaryotes from prokaryotes, which later subdivided into archaebacteria and eubacteria. Hence the rooting problem. In possibilities 1 and 2, the common ancestor must have been a prokaryote, ancestral to eukaryotes by way of archaebacteria in the first case and by way of eubacteria in the second. In the third possibility, the common ancestor could have been anything between a eukaryote and a prokaryote. Unfortunately, available sequencing data not only have failed to provide an unambiguous answer to the rooting problem but have yielded conflicting answers. These questions are highly technical and concern the interpretation of the data at least as much as the data themselves. To sum up a complex situation, most authors favor a prokaryotic root. Woese’s proposal of a common ancestor related to the most thermophilic (heat-loving) archaebacteria, which sequencing has identified as particularly ancient, is widely accepted. Even an ancestor more closely resembling eubacteria is believed to have been adapted to a high temperature, since the most thermophilic eubacteria are also the most ancient of the group.4 The origin of the eukaryotic line remains uncertain. Eukaryotes share many properties with archae¬ bacteria, but also a few with eubacteria. I shall examine in chapter 14 the various explanations that have been proffered to account for these discrepancies. Not all researchers accept a thermophilic prokaryotic ancestor. The French investigator Patrick Forterre5 has argued strongly that thermophily cannot go back to the common ancestor because a primitive system is unlikely to have withstood the harsh conditions imposed by a very high temperature. He believes instead that adaptation to a high temperature is a later development that was achieved through simplification. According to Forterre, the common ancestor was a primitive eukary¬ ote, and prokaryotes were born by a streamlining process associated with the inva¬ sion of an increasingly hot environment. The prokaryotic way of life, instead of being primitive, as generally assumed, would thus have arisen as a secondary adap¬ tation to heat. Once it had arisen, this type of organization would have proved enor¬ mously successful and invaded all niches presently occupied by bacteria. An even stranger hypothesis has been advanced by Mitchell Sogin,6 an expert in

THE ANCESTOR OF ALL LIFE

115

the field of comparative sequencing from the Marine Biological Laboratory at Woods Hole, Massachusetts. According to Sogin, the common ancestor was a prim¬ itive cell—a progenote, to use a term coined by Woese—straight out of the RNA world, and DNA appeared only after the first forking of the tree of life, in the line that was to lead to prokaryotes. The RNA line went on developing into a large cell resembling eukaryotes in several respects, but devoid of a nucleus and lacking the entire machinery involved in DNA synthesis, replication, and transcription. This cell then allegedly acquired a nucleus and its machinery by engulfing a prokaryote, probably of archaebacterial lineage. I shall come back to these proposals later, when examining the early evolution of prokaryotes and eukaryotes. For the time being, I shall stick to the commonly accepted hypothesis.

PORTRAIT OF AN ANCESTOR The common ancestor of all life was a single-celled organism of prokaryotic type, that is, resembling present-day bacteria in lacking a fenced-off nucleus and having only a rudimentary internal organization. This appears as the more probable possi¬ bility in the present state of our knowledge, although alternative descriptions have been proposed. What did this organism look like? There is no clue to this question. The familiar picture of the blunt, rod-shaped Escherichia coli, the main resident of our gut and the most extensively studied specimen of the bacterial world, is a misleading stereotype. Bacteria come in all shapes—spherical, cylindrical, and filamentous. Some microbes recently isolated from deep-sea hydrothermal vents even look for all the world like miniature, flat, sharp-edged rectangular tiles. Some ancient microfossils are long, thin, threadlike structures. But this is no more than a hint. The choice of ancestral shape is entirely open. Although rare wall-less bacteria exist today, it seems likely that the ancestral cell was surrounded by a solid wall. Wall-less forms are very fragile, and the ancestral cell probably could not have survived without some outer protection. Also, we know from microfossils that organisms encased by a wall existed as early as 3.5 bil¬ lion years ago. In all likelihood, the plasma membrane of the ancestral cell was built on the uni¬ versal theme of the lipid bilayer with inserted transmembrane proteins. Granting the existence of a typical membrane, the question arises as to which of the many specialized systems mentioned in the preceding chapters were already incorporated within it and which came later. A valuable clue is provided by the fact that the ancestral cell almost certainly used protonmotive force. This major mechanism of energy retrieval is too wide¬ spread to be a later product of evolution. So are several of the main components

116

THE AGE OF THE PROTOCELL

of membrane-bound electron-transfer chains, including iron-sulfur proteins, hemoproteins, flavoproteins, and perhaps others. The most elaborate respiratory chains were probably yet to come, but some of their main components were already in place. On the other hand, it is likely, although this point is debated, that the ances¬ tral cell lacked the ability to use light energy for the generation of protonmotive force. Be that as it may, the use of protonmotive force indicates that the ancestral cell membrane was impermeable to protons and other ions, and therefore to most of the molecules that must move in and out to satisfy the metabolic requirements of a cell. Therefore, the membrane must have possessed the minimum number of transport systems needed for metabolic exchanges with the environment, and these systems must have been sufficiently sophisticated to operate without letting protons through. The ancestral membrane must also have possessed the various enzymes and insertion systems needed for its own construction. In addition, the membrane must have included the translocation systems needed for the extrusion and assembly of the surrounding wall constituents. Quite possibly, the ancestral cell was able to secrete proteins and carried out extracellular digestion. The widespread distribution of the mechanisms involved in these activities and their close molecular similarities throughout the living world strongly support these contentions. We do not know to what extent the ancestral cell was equipped with surface receptors, or whether it possessed motile or sensing structures, including sex pili. These possibilities are by no means excluded. There can be little doubt that the ancestral plasma membrane and its associated elements already displayed many of the structural and functional attributes that characterize bacterial cell membranes today. Metabolically, the ancestral cell carried out all the reactions needed for the con¬ struction and breakdown of its constituent molecules and for the support of its energy requirements. It did so by proven pathways now operative in a wide variety of prokaryotic and eukaryotic organisms. It had available for this purpose many of the coenzymes found in present-day cells, and used ATP as the main purveyor of energy. Some details of the ancestral metabolism must remain conjectural, as they depend on the kind of environment the cells occupied. We have seen that because the most an¬ cient among known bacteria live in a hot environment, it is widely believed, though not unanimously, that the common ancestor occupied such an environment. Like present-day thermophilic organisms, it may have used some sulfur compounds as a final electron acceptor, or perhaps ferric iron, as I have suggested. The question has often been raised whether the ancestral cell fed on pre-existing organic molecules, as heterotrophs do today, or shared with extant autotrophs the ability to build organic molecules from simple mineral precursors. The heterotroph theory was popular in the days when the ancestral cell was viewed as a resident of the primeval soup. Indeed, such must have been the case for the first protocell. The ancestral cell, however, is the product of a long evolutionary history, in the course

THE ANCESTOR OF ALL LIFE

117

of which complex electron-transfer chains were assembled, and autotrophy, there¬ fore, could very well have developed. Also, the abiotic supply of organic molecules must have dwindled progressively and could well have died out long before the ancestral cell appeared. Finally, autotrophy is sufficiently widespread to allow the hypothesis that it was already an attribute of the ancestral cell. I consider it likely, therefore, that the ancestral cell was autotrophic, though not necessarily phototrophic. It more probably resembled present-day chemoautotrophs and, like these organisms, relied on mineral electron-transfer reactions to satisfy its requirements for energy and high-grade electrons, though not with oxygen as the final electron acceptor, but rather ferric iron or some other mineral substance. Last but not least, it is highly likely that the ancestral cell had DNA genes, prob¬ ably strung together in a single, circular chromosome; that it transcribed these genes into RNA; and that it translated most of these transcripts (except those that served catalytically or structurally) into proteins according to the universal genetic code. The ancestral genes and the corresponding proteins had reached the length and complexity that they have in all living organisms today. The proteins were assembled on typical ribosomes with the help of the complex machinery that, with minor variations, is universally operative. In conclusion, we may take it that the ancestral cell was a fairly typical prokary¬ ote, which we might well mistake for some contemporary bacterium if we should happen to meet it today. Exactly which kind of modern prokaryote the ancestral cell would have resembled most is, however, beyond our knowledge, because several blanks unfortunately remain in the portrait we are able to draw. Following are some of the major questions awaiting a clear answer: Was the ancestral cell wall made of murein or other constituents? Were the ancestral membrane lipids of the type chemists define as ether lipids, or were they of the ester type? Was the ancestral organism phototrophic or simply chemoautotrophic? Were the genes of the ances¬ tral cell split by introns or were they continuous? These uncertainties exist because we are faced in each case with clear-cut differences, obviously of very ancient ori¬ gin, among major groups of extant organisms. I shall address these questions in due time when examining the early evolutionary history of the ancestral cell’s progeny. Some readers might be dismayed by the many blanks in the portrait I have painted. I would rather have them marvel at the details that have been gathered—all in the lifetime of this writer—about a minuscule entity of enormous complexity that existed some 50 million human lifetimes ago.

Chapter 13

The Universality of Life

All extant living organisms are descendants of a single ancestral form of

life. So much is clear. But why is it so? There are several possible answers to this question. First, we could, with the proponents of an extraterrestrial origin of life, identify the ancestor with the immigrant germ that seeded the Earth four billion years ago. A second explanation is that no other ancestral form was possible. It is single because it is unique. A third possibility is that the ancestral form arose among a number of competing forms by a process of Darwinian selection. Or, alternatively, there were several forms to start with, but all the other lines are extinct. Finally, there is the possibility that a mere accident caused the ancestral form to emerge among several that were equally possible. Having agreed to disregard the first possibility, we are left with the other four. The central issue here is the old dichotomy between chance and necessity. How much in the common ancestor was due to contingency, how much to determinism? We have no solid clues to answer this question, only surmises based on what we know of the nature of life, and suspect of its origin.

IS LIFE UNIQUE? Given the physical-chemical conditions that prevailed on our planet 3.8 billion years ago, a protometabolism leading to RNA-like molecules was bound to arise along well-defined, reproducible chemical lines. Such is the unambiguous conclu¬ sion I have drawn from a consideration of the mechanisms involved. Because of the congruence rule, this conclusion extends to all features of today’s metabolism that were prefigured in protometabolism, including such key elements as electron trans-

THE UNIVERSALITY OF LIFE

119

fer, group transfer, thioester-dependent substrate-level phosphorylation, the central role of pyrophosphate bonds with, most likely, a privileged position for ATP, and perhaps the participation of several major coenzymes, such as pantetheine phos¬ phate, coenzyme A, and NAD. Life is very much constrained by its early chemistry, which was itself ruled by deterministic factors. What about other kinds of “life,” based on a different kind of chemistry, bom under different physical-chemical conditions, and adapted to a different environ¬ ment? I cannot reject such possibilities outright, but I find it unprofitable to raise them on the pretext of leaving no stone unturned, as long as not the slightest clue to their reality, or even plausibility, is available. The properties we most intimately associate with the concept of life depend on versatile macromolecules that all chemists agree could not be built with other than carbon frameworks. Even silicon, carbon’s closest relative, would not do. Water is uniquely suited as a medium for life. No other liquid is known that has a comparable combination of favorable physical properties. In addition, water provides two indispensable elements for the construction of carbon-containing molecules, hydrogen and oxygen. The irreplaceability for life of nitrogen, sulfur, phosphorus, and other biogenic elements has like¬ wise been emphasized. Add to these considerations the predilection of interstellar chemistry for compounds of these elements and you end up with a strong case in favor of a life uniquely constructed according to the same kind of “organic” chem¬ istry. Whether this kind of chemistry could, under a different set of conditions, develop into life-generating worlds different from the RNA world is a question that must be left open. Much contemporary research in organic chemistry aims at “imi¬ tating” life with artificial molecules. Even if such efforts should one day be suc¬ cessful, the question of the artificial process ever occurring under natural condi¬ tions would still have to be answered. Until this happens, if it ever does, let us be content with the life we know. It is wondrous enough to make recourse to other hypothetical lives unnecessary. Within the straightjacket of its chemistry, could emerging life have evolved a different genetic system? I addressed this question in part II and reached the con¬ clusion that only secondary details—perhaps in the genetic code, although even that is far from certain—could have been different. Contingency no doubt played a role in shaping the exact evolutionary history of the RNA world, but the stringency of the selection factors ensured that the end result could hardly have been different, including the formation of protocells, a mandatory condition of further evolution at a certain stage. There remains the long pathway from protocell to common ancestor. Chance played a role at every step of the pathway by providing an appropriate mutation, thus offering enormous scope for diversity. There was a major bottleneck, however: the need to develop autotrophy before the supply of abiotic products became exhausted. By that time, any heterotrophic line that might have existed was per¬ force extinct. The survivors either had available a direct supply of electrons at level

120

THE AGE OF THE PROTOCELL

A from some mineral source—a very unlikely occurrence according to our knowl¬ edge of the mineral world—or had converted their thioester-dependent machinery to the reverse transfer of electrons from level B to level A, with the help of energy supplied by the hydrolysis of ATP. This implies that they had developed an alterna¬ tive machinery for generating ATP and had available an appropriate mineral source of electrons with, if necessary, the means of bringing these electrons to level B. These conditions may have left little leeway for contingency, especially if, as I suspect, circumstances—such as a high environmental acidity—put a high selective premium on the development of energy-driven proton extrusion. Under such cir¬ cumstances, harnessing protonmotive force could well have been the only means of getting through the autotrophy bottleneck. If not, it was most likely the most effi¬ cient and, perhaps, the easiest to attain, considering that the required cofactors, such as the flavin derivatives FMN and FAD and, perhaps, even porphyrins may have been present as products of the carbon-nitrogen combinatorial chemistry that generated the RNA world. There may have been some competition at the entrance of the bottleneck, but little choice at the exit. If my reconstruction is correct, there¬ fore, some cell resembling the common ancestor in its main characteristics may well have been the obligatory outcome of the biogenic process set off on the Earth some 3.8 billion years ago. In the introduction to this book, I argued on theoretical grounds—remember the thirteen spades—that the emergence of life must have involved a very large number of steps, most of which had a high probability of occurring under the prevailing conditions. But I left open the possibility that there might be more than one path¬ way compatible with this exigency. My conclusion, after a consideration of the underlying chemistry, is that, given the opportunity, the development of life is very likely to take the course it actually took, at least in all essential aspects.

EXTRATERRESTRIAL LIFE Is there life elsewhere in the universe?1 The two Viking spacecraft (launched in 1976) carried equipment designed to answer this question by probing for traces of life on Mars. The results, unfortunately, were negative or at best “ambiguous.” But Mars is only our nearest neighbor. What about other solar systems? There are about 100 billion stars in our galaxy alone, and there are billions of galaxies in the uni¬ verse. How many of the trillions of existing stars have planets? How many of those planets would have a geological history comparable to that of planet Earth? On how many of those would the physical-chemical setting that fostered the birth of life on Earth be duplicated? And, finally, on how many of the planets exhibiting these conditions would life in fact emerge, and how closely would that life resem¬ ble life on Earth? Nobody knows the answers to these questions, but they have been much in the

THE UNIVERSALITY OF LIFE

121

limelight since a memorable day, November 1, 1961, when a number of scientists met at the National Radio Astronomy Observatory in Green Bank, West Virginia, to launch the search for extraterrestrial intelligence (SETI). On this occasion, they devised the so-called Green Bank equation, which includes as relevant parameters the number of stars capable of having planets in the universe and the probability of a planet’s being capable of harboring life. Some of the greatest experts on cosmol¬ ogy have pondered these problems in the light of available evidence. Although quantitative estimates vary widely, the consensus is that the history of the Earth is probably not unique. The figure of about one million “habitable” planets per galaxy is considered not unreasonable. Even if this value were overestimated by several orders of magnitude, it would still add up to trillions of potential cradles for life. If my reading of the evidence is correct, this means that trillions of planets exist that have borne, bear, or will bear life. The universe is awash with life. Unfortunately, interstellar distances are such that we may never get confirmation of this, unless extraterrestrial life somewhere reached a stage of development that rendered it capable of sending out messages that we can receive and decode. Hence the interest of the Green Bank participants in the search for extraterrestrial intelli¬ gence. In the meantime, we must rest content with the message we get from life down here, and that message is that there must be plenty of life out there. *

ARTIFICIAL LIFE This term has nothing to do with life in a test tube, although this may happen some day, but rather with life in a computer.2 Ever since the famous Hungarian-American mathematician Johannes von Neumann devised the first “cellular automaton” in the late 1940s, theoreticians have evinced considerable interest in the mathematical modeling of such typical properties of living organisms as complexity, self-organi¬ zation, development, reproduction, and evolution. They have paid special attention to the spontaneous emergence of these properties as a result of interactions among different variables intended to represent catalysts and their reagents, or genes and their products. These models have highlighted conditions under which order can arise out of disorder through fluctuations that take place randomly until the system becomes caught in a network of interactions that drives it toward a dynamically organized configuration in which it settles. What is depicted is the stochastic exploration of a “space,” eventually leading the system to fall into a “basin.” Stuart Kauffman,3 a pioneer in the field and a prominent member of the Santa Fe Institute, which has become the mecca of artificial life, prefers the opposite images of a “rugged fitness landscape” with “adaptive peaks” separated by “valleys.” Somehow, I find the image of falling into a basin more representative of reality than that ot climbing to the top of a peak.

122

THE AGE OF THE PROTOCELL

Irrespective of the imagery, what the new computing methodologies have revealed is how a system of multiple interacting variables may become stranded in a basin (or on a peak), and how, depending on the structure of the landscape, it may escape and land in another basin (or on another peak), according to a saltatory, non¬ linear kind of process, similar in its mode of unfolding to what evolutionists call punctuated equilibrium. This process includes prolonged periods of pseudostability during which little changes connected to each other by occasional jumps induced by some chance event. According to Kauffman, the condition for this process is a kind of restricted instability separating chaos from fixity. “Life,” he concludes, borrowing a phrase from his colleagues Norman Packard and Christopher Langton, “adapts to the edge of chaos.”4 Darwinian evolution operates within the confines of a landscape and can be understood only in relation to the configuration of this landscape. “Artificial life” studies fit within the current interest in dissipative structures, complexity, chaos, catastrophe, turbulence, and other phenomena that obey non¬ linear relationships such that very small changes may precipitate major events, the so-called butterfly effect: A butterfly fluttering in Rio unleashes a storm over Chicago. Many natural phenomena owe their relative unpredictability to this kind of intermingling of stochastic and deterministic factors. Viewed in a certain per¬ spective, life would seem to be a particularly striking example of such an occur¬ rence. The possibility that the spontaneous origin and development of life could be accounted for through the “freezing” of some fortuitous configuration of matter has obvious appeal, especially among those who reject deterministic explanations. The modeling approach, which uses elegant mathematical procedures, has thrown much valuable light on the intrinsic properties of self-organization and self¬ regulation that characterize all living systems. It has also illuminated some impor¬ tant aspects of biological evolution. But the term “artificial life,” applied by anal¬ ogy with “artificial intelligence,” could be misleading. Life is a chemical process. If it is ever to be created artificially, it will be by a chemist, not by a computer.

PART IV

THE . AGE OF THE SINGLE CELL

Chapter 11+

Bacteria Conquer the World

The common ancestor of all living things most likely was a bacterium, or

prokaryote. Were it not for one line—which entered the long, complex, and myste¬ rious pathway that led to eukaryotes—all of its progeny today would consist exclu¬ sively of bacteria. Even though they are no longer alone, bacteria still make up the larger part of the living world. Their evolution from the common ancestor illus¬ trates the astonishing durability and versatility of prokaryotic forms of life. These qualities have allowed them to adapt to all sorts of different environments and to establish themselves and flourish in almost every kind of habitat. The diversity of bacteria is staggering and still incompletely inventoried. The reason for this success is simple: Bacteria are built to grow and multiply as fast as materially possible. They epitomize life at its rawest, with no frills.

THE SECRET OF BACTERIAL SUCCESS A bioengineer attempting to construct a cell designed to proliferate as fast as possi¬ ble could not come up with anything better than a bacterial cell. The bacterial genome is “streamlined” for fast replication. Genes are not split by introns and are crammed in the chromosome with hardly any space left for “junk” DNA. The chro¬ mosome itself is loosely structured, offering little impediment to the replication process. Furthermore, bacteria hardly ever stop duplicating their DNA and they manage to transcribe their genes and build all the RNAs and proteins they need for growth while they go about this activity. Some even start a second round of dupli¬ cation before the first one is finished. As soon as two copies of their genome are available, they divide. As a result, it takes the average bacterium no more than twenty to thirty minutes to go through a complete growth and division cycle, as opposed to some twenty hours for the average animal or plant cell.

126

THE AGE OF THE SINGLE CELL

The bioengineer responsible for this feat of design was natural selection. We are therefore led to ask what evolutionary advantage could have driven the process. It cannot have been mere rapidity in producing progeny. A single bacterial cell would, by unrestricted exponential growth, cover the whole surface of the Earth with off¬ spring in less than two days. It would take a eukaryotic cell little more than two months to achieve the same result. Clearly, lack of available resources will soon curtail multiplication in either case. There seems little advantage in beating the gen¬ eration clock. No, the main advantage bacteria gain from their fast multiplication rate lies in the enormous number of mutants they can offer to natural selection. By the time two eukaryotic cells have arisen out of one, a bacterial cell can produce up to one trillion cells, among which several billion mutations due to replication errors alone would be distributed. (A bacterial genome contains about three million base pairs, and the minimum replication-error frequency is of the order of one wrongly inserted base in one billion.) Many of these mutations will be neutral, that is, they will have no effect on the proliferating ability of the cell. Many others will be dele¬ terious and will be weeded out by natural selection because the affected cells can¬ not multiply, or multiply more slowly than the others. But the odds are that an occa¬ sional mutation will turn out to be advantageous, especially if the environment changes. That is why the fight against disease-causing microbes never ends. What¬ ever new antibiotics may be discovered, some resistant mutant is likely to arise and to proliferate preferentially in the presence of the drug. This no doubt happened to bacteria countless times during their long evolution. Whenever conditions changed, some mutant was there that could take advantage of the new conditions. This versa¬ tility has allowed bacteria to invade every possible ecological niche. Bacteria do indeed cover the whole surface of the Earth with multiple layers of thriving life. They are the great survivors. The evolutionary strategy of bacteria illustrates a statistical form of adaptability that is foreign to our intuitive understanding of this term. We tend to think of adap¬ tation in terms of individual responses—conscious and deliberate or unconscious and automatic—to changing circumstances. We react to cold by putting on more clothes. The pupils in our eyes react to strong light by contracting. If we were bac¬ teria, the individual would not count. We would let most of us perish frozen or blind, and we would rely on the odd person who happened to be warmly covered or to have naturally narrow pupils to rapidly replenish our stock with similarly adapted individuals. Should the temperature go up again or the light diminish, the same strategy would quickly bring back a population of scantily clad or wide-eyed individuals. Human beings could not behave that way, even if unhampered by any sort of respect for the individual, because it would take them too much time to recover. Bacteria can behave that way because of their rapid proliferation rate. Note, however, that the bacterial strategy has also been followed by more complex organisms, but much more slowly and gradually, over eons of time; variation screened by selection is the mainspring of Darwinian evolution.

BACTERIA CONQUER THE WORLD

127

Bacteria do have some built-in mechanisms for individual adaptability. For example, certain bacteria, when transplanted into a medium containing galactose (milk sugar) as the sole source of carbon, respond by producing the enzymes they need to utilize this sugar. This is a famous case in the history of science because it suggested at first that galactose somehow “instructs” the cells to make the appropri¬ ate enzymes. The French investigators Francis Jacob and Jacques Monod gained a Nobel Prize for demonstrating that the bacteria have the ability to make the enzymes all the time but fail to make them in the absence of the sugar because the corresponding genes are blocked by a protein, called a repressor, that prevents their transcription into RNA. A derivative of the sugar unlocks the genes by binding to the repressor in shell a way that the repressor can no longer exert its blocking action.1 This mechanism is understandable in the context of the general streamlin¬ ing evolutionary strategy of bacteria. Don’t waste time and energy making some¬ thing until it is needed. The operation of the immune system in humans and higher animals offers another intriguing example of an apparently instructional mechanism. When the organism is exposed to a foreign macromolecule, known as an antigen, it responds by making the corresponding antibodies, which are proteins that specifically neutralize the antigen. If the antigen is borne by a virus, a microbe, or a foreign cell, from a transplant, for example, the response includes the development of killer cells specific for that par¬ ticular antigen. Cancer patients may even mount an attack against their own tumor cells in this manner. The cells responsible for manufacturing antibodies and killing foreign cells are called lymphocytes. They are generated in the bone marrow and cir¬ culate in the blood. When, in the early 1950s, these mechanisms began to be eluci¬ dated, it seemed obvious that the antigens must be instructing the lymphocytes. How else could one explain the fact that virtually any antigen could elicit a specific re¬ sponse? Only the Australian immunologist Macfarlane Burnet thought otherwise. He proposed the view, which appeared fantastic at the time but turned out to be correct, that small numbers of lymphocytes capable of fighting almost every possible kind of antigen exist preformed in the organism and that when any such cells come into con¬ tact with “their” antigen, they proliferate to form a clone (a population of identical cells) of antibody-making or killer cells specifically shaped to recognize the foreign antigen.2 Known as clonal selection, this mechanism recalls the bacterial strategy of hav¬ ing mutants waiting in the wings, so to speak, for a wide variety of occasions. But it is of a much higher degree of sophistication. The lymphocyte “mutations” are not accidental but programmed. During their maturation, lymphocytes undergo com¬ plex genetic rearrangements that create the millions of different genes—perhaps as many as one billion—responsible for their diversity. In this mechanism, which recalls the primeval modular game whereby the first large genes were assembled, a piece of the gene—let us call this piece A—is taken at random from a set of differ¬ ent A pieces and joined with B, C, D, and E pieces similarly taken from correspond¬ ing sets. Thanks to this process and to some additional causes of diversification,

128

THE AGE OF THE SINGLE CELL

each lymphocyte ends up with a different ABCDE combination, which is translated into an antigen-recognizing protein of different specificity. Also much more elaborate than the passive proliferation of the better-adapted bacteria is the mechanism whereby lymphocytes are specifically triggered to multi¬ ply upon contact with the appropriate antigen. The whole problem of growth con¬ trol—and of its derangement in cancer—is posed here. Once we have stopped growing, most of our cells no longer divide. Exceptions include the cells that must be renewed because they have a short life span, as do many blood cells, or because they are lost by sloughing off, as are the cells that line our skin and mucosae. Most of our cells do, however, retain the capacity to multiply when appropriately trig¬ gered, for example, in wound repair. Enormous advances have been made in recent years in our understanding of these mechanisms, which involve a number of recep¬ tors and cell type-specific “growth factors” that activate the receptors. Many can¬ cerous transformations are due to alterations of one of the genes that code for these receptors and growth factors. Known as oncogenes (cancer genes) for this reason, these genes all belong to the normal growth-control machinery. In the clonal selection of lymphocytes, exposure to a given antigen selectively triggers the division of the cells that recognize the antigen, thus building an army specifically directed against the enemy. There is a disadvantage to all this sophisti¬ cation, namely, slow proliferation. It takes lymphocytes a couple of weeks to build a clone that bacteria would generate in a few hours. In addition, there is the danger of derailment of the growth-controlling machinery, leading to such deadly diseases as lymphoma and leukemia, which are “lymphocyte cancers.” Even though modern bioengineering may not have invented the streamlined, ultrafast multiplier, it is now making full use of it. In gene cloning, a piece of for¬ eign DNA is introduced into a bacterial host by genetic-engineering techniques, and the engineered cell is then given the opportunity for unrestricted clonal growth. The next day, appreciable amounts of the inserted gene, which was duplicated with the bacterial DNA at each generation, are available for sequencing and other uses. If the gene has been inserted in such a way that it is actually expressed, its product can similarly be obtained in essentially limitless amounts. This is how human insulin is now being made industrially by bacteria. Even slow multipliers are now used as factories. Lymphocytes rendered able to multiply indefinitely by fusion with a cancer cell are cloned on a large scale for the production of specific anti¬ bodies, called monoclonal for this reason.

THE FIRST FORK Some time around 3.8 to 3.6 billion years ago, some members of the primordial ancestral cell population became separated from the bulk of the group by some geo¬ logical or climatic phenomenon that brought them into a different, less favorable

BACTERIA CONQUER THE WORLD

129

environment. This would have been their undoing but for a rare mutant—remember the bacteria exposed to an antibiotic—that happened to be adapted to the new cir¬ cumstances and proliferated. We know this because both the parent line and the off¬ shoot have grown, evolved^ and diversified into a wealth of different varieties that are now found, often side by side, in every part of the world. Yet unmistakable proof of their kinship within one or the other group—either archaebacteria or eubacteria—and of the deep, primeval rift between the two groups remains inscribed in the structures of some key constituents, in certain characteristic meta¬ bolic reactions, and, especially, in the sequences of nucleic acids and proteins. And so, amazingly, events that happened eons ago in microscopic entities that have left no decipherable trace in the fossil record have been uncovered and reconstituted in some detail thanks to the magic of modem molecular biology. What was the environmental change that triggered the rift? We don’t know but we can hazard a guess. If, as is likely, though not certain, the universal ancestor was a prokaryote of archaebacterial type adapted to a high temperature, a plausible hypothesis is that the dissident group found itself in a cooler environment where heat resistance turned into an impediment. There is some support for this possibil¬ ity. Whereas archaebacteria exist that thrive at temperatures of up to 110°C (230°F)—with pressure sufficient to prevent the water from boiling—no ther¬ mophilic eubacteria are found at a temperature higher than 80°C (176°F), and, according to comparative sequencing, those that live at this temperature are among the most ancient. If this hypothesis is correct, what aspect of heat adaptation would have become unfavorable to survival in a cooler environment and could have been offset by an appropriate mutation? Extremely thermophilic archaebacteria survive and multiply in their inhospitable habitat thanks to their possession of heat-resistant proteins and other constituents. Most proteins unfold and lose their specific conformation irre¬ versibly when heated to temperatures on the order of 50° to 70°C (122° to 158°F). The coagulation of egg white is an example. The usual enzymes become inactive under such conditions. The proteins of thermophilic organisms are much more resistant to heat denaturation, a property that is now attracting great interest on the part of those who want to use enzymes as industrial catalysts. Also characteristic of thermophilic archaebacteria,3 of all archaebacteria, in fact, are special membrane lipids, known as ether lipids, that form particularly strong bilayers. (Ethers arise from the joining, with loss of water, of an alcohol molecule with another alcohol molecule.) In the most extreme thermophiles, the bilayer is further welded into a rigid structure because the hydrophobic ends of the lipids are joined chemically into single chains. In eubacterial membrane lipids, the rigid ether bond is replaced by the more flexible ester bond (between an alcohol molecule and an acid molecule), and the two layers of the bilayer can slide freely. Consider now what could have happened to highly thermophilic bacteria sud¬ denly exposed to a (relatively) cooler environment. The possession of heat-resistant proteins could hardly have been a drawback, especially one correctible by a single

130

THE AGE OF THE SINGLE CELL

mutation, unless one specific protein became locked in an unfavorable configura¬ tion below a certain critical temperature. In contrast, the rigid ether lipids could have been a major handicap. In order to understand this, think of lard, butter, and salad oil. Each changes from solid to liquid at a given temperature. Salad oil is liq¬ uid at room temperature but congeals in the refrigerator. Butter is solid but melts in hot weather. Lard melts at an even higher temperature. All three natural fats consist of similar substances known as triglycerides. Secondary chemical differences among the triglycerides account for the differences in melting temperature. The same is true for membrane lipids. Depending on their chemical composi¬ tion, their melting temperatures may differ as much as do those of lard and salad oil. In particular, ether lipids generally melt at higher temperatures than do the cor¬ responding ester lipids. Furthermore, there is a clear correlation between the melt¬ ing temperature of a cell’s membrane lipids and the temperature of the environment occupied by the cell. This is understandable. Membranes can accomplish their functions only if their lipid bilayers are kept fluid. On the other hand, excessive flu¬ idity of the membranes may endanger cellular stability. So, each cell type has mem¬ brane lipids that are fluid at the normal surrounding temperature but congeal about 10° to 15°C (18° to 27°F) below this temperature. With this information, it is easy to visualize what would have happened to the highly thermophilic archaebacterial ancestors transferred by some climatic or geo¬ logical change from, say, 110°C to 80°C. Their ether membrane lipids would con¬ geal, the cells would become sluggish, their exchanges with the environment would come to a halt, and the cells would literally freeze to death, albeit in water we might still consider scalding. Only a mutation leading to the formation of membrane lipids that remain fluid in the new environment could save the cells from this sorry fate. This is what I suggest may have happened, the mutation being one that replaced ether lipids by ester lipids in the mutant cell’s membranes. The price the mutant cell and its progeny paid for their rescue was that they could no longer return to their boiling cradle. But the bounty was immensely greater. The whole world was theirs to invade. The first eubacteria were born. This story of the genesis of eubacteria is hypothetical. The alternative possibil¬ ity, that ester lipids were inherited from the common ancestor and ether lipids acquired by archaebacteria, cannot be excluded. My choice is based on the assump¬ tion that the ancestral cell occupied a hot environment and possessed the betteradapted lipids. This view is shared by many scientists, but not by all. Some members of the eubacterial family have kept a predilection for a hot envi¬ ronment—though not as hot as that occupied by the most thermophilic archaebacte¬ ria—or have returned to such an environment at a later evolutionary stage. You can see some at work creating their favorite surroundings by their own metabolism in the steaming compost heap at the back of your garden. Most eubacteria, however, are adapted to milder temperatures. Some even thrive in the icy waters around the polar caps. In contrast to archaebacteria, which, with rare exceptions, have remained confined to their original hot niches and to a few specialized environ-

BACTERIA CONQUER THE WORLD

131

merits, eubacteria are found everywhere; they are by far the dominant prokaryotes. Among them are all the bacteria that cause diseases and the many additional harm¬ less ones that we harbor in our gut and elsewhere on our body, but also a multitude of other invisible organisms that reveal their existence by fermentation, food spoilage, rotting of organic matter, and other natural manifestations. Fungi, which are eukaryotes, also play an important role in these processes.

OUTLANDISH COLONIES *

While eubacteria were conquering the world, the archaebacteria may have long remained confined to the boiling waters of their birth, to which they were particu¬ larly well adapted. At this stage, these organisms presumably lost the ability to syn¬ thesize the characteristic cell-wall constituent murein, which is found exclusively in eubacteria. Murein, with its content of both D- and L-amino acids and its other structural irregularities, has the characteristics of a primeval substance. Therefore, the ability to make murein was more likely lost by archaebacteria than acquired by eubacteria. Most archaebacteria do have a cell wall, but made of protein and carbo¬ hydrate constituents different from murein. Eventually, some archaebacteria ventured outside their original medium and succeeded in invading other habitats.4 A particularly flourishing group developed— or quite possibly retained, as we may be dealing with a primitive metabolic attribute—the ability to use hydrogen for the anaerobic conversion of carbon diox¬ ide to methane, and to support all their energy requirements with the help of this reaction. Methane is a highly flammable substance, the most volatile component of natural gas. In line with their presumptive origin, the most ancient methanogens, as these organisms are called, are thermophilic. Later forms acquired the capacity to grow at lower temperatures while retaining ether lipids in their membranes. They are now established in almost every site where organic matter is decomposed anaerobically with the generation of hydrogen. They are found in the digestive tracts of animals, especially of cattle, which have become significant producers of atmospheric methane and, thereby, participants in the greenhouse effect (see chapter 30). Methanogens are also abundant in marine and freshwater sediments. From such muddy depths, they send up the bubbles that break the silence of swamps by their muffled plopping, and fuel the will-o’-the-wisps that flit on the surfaces of marshes at night. Other archaebacteria have succeeded in colonizing waters of very high salinity, even the saturated brine of drying seas. They are the only living things that still inhabit the Dead Sea and the Great Salt Lake. Among these remarkable salt-loving organisms, or halophiles, is found the only known archaebacterial phototroph: Halobacterium halobium. Unlike all other phototrophs, this organism does not

132

THE AGE OF THE SINGLE CELL

depend on chlorophyll to catch light. Instead, it relies on a purple substance known as bacteriorhodopsin. This substance is a membrane-bound protein linked with a carotenoid, a relative of vitamin A, which acts as the light-catching part of the com¬ plex. Carotenoids are found everywhere in the living world, including the specialized membranes of phototrophic organisms. But bacteriorhodopsin offers the only known example of a substance of this family actually effecting the conversion of light into usable energy. This example is also unique in another respect. The absorbed light is used directly to generate protonmotive force, without the partici¬ pation of electrons. Unlike chlorophyll, bacteriorhodopsin is not a light-powered electron pump; it is a light-powered proton pump. It is interesting, and possibly revealing, that the closest chemical relative of bac¬ teriorhodopsin is the light-sensitive purple pigment of animal eyes. This is the orig¬ inal rhodopsin, a name that combines the Greek roots for rose and vision. Because of this relationship with the eye, carotenoids are also known as retinoids. In vision, however, the excited rhodopsin does not fuel an energy-converting mechanism; it triggers a series of signals along nerves leading from eye to brain. It is, however, tempting to assume that this vital pigment of our eyes is a descendant of some remote, ancestral bacteriorhodopsin.

THE GREEN REVOLUTION In chapter 10, I related how some red cytochrome turned into a green chlorophyll through the evolutionary appearance of a variant porphyrin molecule in which the central hole came to be occupied by a magnesium atom in lieu of an iron atom. This event most likely took place in the eubacterial line after this line separated from the archaebacterial line, as no organism endowed with chlorophyll is known among archaebacteria and only a small number of eubacterial species are phototrophic. It is tor these reasons that I have assumed that the ancestral cell was not phototrophic. The alternative possibility, that the ability to make chlorophyll is an ancestral heir¬ loom that was lost by all archaebacteria and by many eubacteria, seems much less likely. The appearance of chlorophyll had consequences of the utmost importance, first for the bacteria concerned and, eventually, for the whole living world and for the planet itself. As mentioned earlier, photosystem I appeared first in the conquest of solar power by emerging life. When energized by light, photosystem I can draw electrons from mineral compounds, sometimes also from organic substances, but not from water. A number of phototrophic bacteria dependent on photosystem I are found in the world today. The next major step in the green revolution was the development of the water-

BACTERIA CONQUER THE WORLD

133

consuming, oxygen-producing photosystem II, presumably through evolutionary modifications of photosystem I. The two systems operate with related but different chlorophylls. In present-day phototrophs, photosystem II is always associated with photosystem I, which gives tfie electrons lifted from water by photosystem II the extra boost they need to participate in biosynthetic reductions. In the bacterial world, the association of the two photosystems is found in a large and abundantly distributed class of bluish microorganisms, originally called blue-green algae because they tend to associate into multicellular chains resembling some primitive seaweeds (algae). However, authentic algae are eukaryotic organ¬ isms. To avoid confusion, the prokaryotic phototrophs possessing the two photo¬ systems are now called cyanobacteria (from the Greek kyanos, blue). When did these important events take place? According to many experts, at least 3.5 billion years ago, perhaps as early as 3.75 billion years ago. The most solid evi¬ dence comes from stromatolites,5 those layered rocks that originate from superim¬ posed bacterial colonies. In extant colonies of this type, the top layers are occupied by cyanobacteria, which serve as essential food providers for the deeper heterotrophic layers. The oldest known stromatolites date back 3.5 billion years. If the bacterial colonies from which these rocks arose were anything like their presentday counterparts, they were topped by cyanobacteria-like organisms, meaning that photosystem II, as well as photosystem I, is at least 3.5 billion years old. This estimate is supported by microfossil traces of the same age. The interna¬ tionally known microfossil expert William Schopf,6 from the University of Califor¬ nia at Los Angeles, has identified authentic traces of at least seven distinct cyanobacteria-like organisms in rocks situated in northwestern Australia and accu¬ rately dated between 3.46 and 3.47 billion years. Many of the traces appear like chains of up to several tens of walled cells almost indistinguishable from the mor¬ phology of some present-day cyanobacteria. If Schopf’s identification is correct—he is the first to admit the uncertainty of purely morphological criteria—the phototrophic production of oxygen started at least 3.5 billion years ago. Yet all the available evidence indicates that molecular oxygen did not start rising in the atmosphere until about 2.0 billion years ago and reached a stable level only around 1.5 billion years ago. A likely explanation of this discrepancy is that, for the first two billion years of oxygen-producing phototrophy, enough oxygen-avid minerals were present to trap the oxygen produced and that atmospheric oxygen started rising only after these oxygen “sinks” were saturated. A major such sink could have been ferrous iron, believed to be very abundant in the early oceans. The reaction of ferrous iron with oxygen could explain at least in part—there are other possibilities—the large-scale generation of the mixed fer¬ rous/ferric oxide deposits (magnetite) that are the main constituents of the banded iron-formations mentioned in chapter 3. It is suggestive and possibly significant that the deposition of banded iron-formations dates back at least 3.75 billion years—the geological record does not go further. It continued uninterrupted until

134

THE AGE OF THE SINGLE CELL

oxygen started appearing in the atmosphere, then declined progressively to come to a halt about 1.7 billion years ago. Note that if banded iron-formations attest to the presence of oxygen-producing phototrophic organisms, the universal ancestor must have arisen some time before 3.75 billion years ago, since one must account for the separation of eubacteria from archaebacteria; for the appearance of chlorophyll in a given kind of eubacteria; and for the evolution of phototrophy to the point where oxygen was produced. Thus the time span left for the emergence of life, up to the common ancestor, would be lim¬ ited to a maximum of some 200 million years, the interval between the time when the Earth first became livable, four billion years ago, and the minimum age of the common ancestor. Such a time span was once considered too short for such a com¬ plex event as the origin of life. As I have pointed out before, there is no valid reason for such a view. Two hundred million years is really an enormous duration, more than twenty times the time it took an ape to become human. During such a stretch, life as we know it could have arisen and disappeared many times. The inauguration of phototrophy was a fateful event in the spread of life on Earth, as it enabled living organisms to plug directly into the huge energy reservoir of the sun for raising electrons to the high energy level required for the construction of biomolecules from mineral building blocks. Earlier forms of life may have done so via the UV-supported production of hydrogen at the expense of the ferrous-ferric conversion described in chapter 3. This latter mechanism had the advantage of sim¬ plicity—an enormous asset at the time life was taking its first faltering steps—but it was no match for the chlorophyll-dependent process once the membrane-embedded infrastructure needed for this process was in place. Another advantage of phototrophy is that it freed autotrophic life from its depen¬ dence on the specific environmental provision of appropriate electron donors and acceptors. Especially after the acquisition of photosystem II, virtually the whole surface of the Earth could be invaded. Our planet became green and its stores of carbon, nitrogen, and other bioelements became increasingly tied up in this green mantle. This situation, in turn, gave a tremendous boost to the development of heterotrophic organisms capable of feeding on the biomolecules manufactured by oth¬ ers. Thus was inaugurated the great coalition that, starting with the first stromatolite colonies, has come to link us, all other animals, all fungi, and many bacteria, to green plants and phototrophic microbes, within a planetary superorganism, the biosphere, whose metabolism is manifested by the continual recycling of the major biogenic elements. Perhaps the most important consequence of phototrophy in relation to evolution was the rise of atmospheric oxygen. The crucial time, in this connection, irrespec¬ tive of the date of the first appearance of photosystem II, was the period between 2.0 and 1.5 billion years ago, which witnessed what was probably the greatest eco¬ logical disaster—and the most far-reaching adaptive reaction of living organisms— in the history of life.

BACTERIA CONQUER THE WORLD

135

THE GREAT OXYGEN CRISIS Until the development of photosystem II, the world had remained virtually oxygenfree. We tend to think of oxygen as a vital element, which indeed it is for us and for all other aerobic organisms, that is, those “living in air.” For the early forms of life, however, oxygen was a redoubtable poison, as it still is for those bacteria known as obligatory anaerobes, which can survive only in the absence of oxygen. The toxic¬ ity of oxygen is due to its ready conversion in the presence of living systems into highly reactive chemical species, with names such as free hydroxyl radicals, super¬ oxide ions, and hydrogen peroxide, that can severely damage vital cell constituents, including DNA and lipid bilayers. When oxygen made its appearance, life had no defense against these poisons, and a major holocaust threatened. Fortunately, the process was slow and there was plenty of time for the main strategies of evolutionary adaptation to come into play. Victims probably were legion, but a few survivors emerged to people the world with new forms of life, thus turning an impending catastrophe into a major source of innovation. The first to adapt to oxygen were its producers. In itself, the key mutation that gave rise to photosystem II was lethal. It may have happened many times, until chance associated it with another genetic change that protected the cells against the toxic effects of the oxygen they generated. This change could have been the ability to make large amounts of substances, called antioxidants or free-radical scavengers, that can mop up the injurious reactive chemicals formed from oxygen. Among such antioxidants are ascorbic acid, or vitamin C, a number of thiols, and tocopherol, or vitamin E. Or the cells could have acquired some protective enzyme, such as super¬ oxide dismutase, which inactivates the superoxide ion, or catalase, which destroys hydrogen peroxide. The need for such protective adaptations could explain the delay that may have separated the appearance of photosystem II from that of photo¬ system I. When other forms of life became exposed to oxygen, their first reaction was retreat. However, retreat from oxygen also meant retreat from the majority of phototrophs, the principal food supply for heterotrophs. In addition, oxygen was all-pervasive. It invaded every crack in the soil and, being soluble in water, reached the depths of the oceans. Sheltered areas capable of supporting anaerobic life soon became scarce. There was thus considerable pressure on existing anaerobes to develop protective mechanisms similar to those that had allowed the original phototrophs to produce oxygen without suffering any harm. Through random muta¬ tions, many bacteria, both autotrophic and heterotrophic, became able to survive in the presence of oxygen. Most did not stop at mere survival. Thanks to relatively simple mutational events, they acquired the ability to transfer to oxygen the electrons exiting from

THE AGE OF THE SINGLE CELL

136

their electron-transfer chains, with the consequent formation of water. The function known today as respiration was thereby initiated. This was a major evolutionary development, which, by returning oxygen to water, inaugurated the planetary water/oxygen cycle. The ubiquitous oxygen replaced the special mineral electron acceptors to which the organisms were chained; electrons flowing through phosphorylating chains could fall right down to the lowest level of energy, thus maxi¬ mizing the potential energy yield of the process. Spurred by these great benefits, many bacteria evolved to the point of becoming obligatorily dependent on their erstwhile foe. Some succeeded in exploiting this dependence to the full by acquir¬ ing respiratory chains that allowed energy retrievals close to the maximum possible yield. Some archaebacteria also acquired the ability to detoxify and exploit oxygen when it started appearing in the atmosphere. The halophilic bacteria are aerobic. So are the thermoacidophilic bacteria, which are residents of what we would view today as the most inhospitable of all ecological niches: waters that are very hot, very acidic, and reeking with the stench of hydrogen sulfide. However, except for its oxygen content, this sort of medium may well have been the cradle of life and the favorite haunt of the universal ancestor. In spite of all these adaptations, oxygen has retained remnants of its life-threat¬ ening properties. Our white blood cells kill microbes with flashes of toxic oxygen derivatives (free radicals). The same derivatives sometimes also form accidentally in our tissues, where they may participate in the aging process, cause genetic lesions, or initiate cancerous transformations. There is considerable interest in the administration of antioxidants as protection against such injuries. The events I have recalled would have been difficult to reconstruct were it not for the rich world of bacteria, where forms representative of almost every interme¬ diate stage in the development of phototrophy and aerobic life can still be found. Their ancientness, as revealed by comparative sequencing and other evidence, gen¬ erally corresponds to their proposed position in the chain of events. But their story goes further. Some of the participants in the great oxygen saga played a fundamen¬ tal role in the remarkable process whereby a prokaryotic precursor turned into the ancestor of all eukaryotes, including ourselves. Thus, our origin, and that of all the plants and animals that surround us, is intimately rooted in the major events of eubacterial evolution, as the next three chapters will show.

Chapter 15

The Making of a Eukaryote

Some 3.5 billion years ago, while bacteria were beginning to cover the world with triumphantly successful colonies, an obscure offshoot started to evolve in a strange direction that would most likely have appeared to an extraterrestrial visitor as totally aberrant in the context of life on Earth as it was then, and leading nowhere. In reality, “nowhere” fanned out, more than two billion years later, into the immensely varied groups of protists, plants, fungi, and animals, including humans—virtually the whole visible part of the biosphere. The way to this extraor¬ dinary diversity of living forms passed through a new type of cell entirely different from any known bacterium, past or present. Called eukaryotic because it possesses an authentic nucleus, this type of cell has many other characteristics that distinguish it clearly from the bacteria, or prokaryotes.

THE PROKARYOTE-EUKARYOTE TRANSITION The nature of the obscure originator of the eukaryotic line is uncertain. Most of the evidence points to a prokaryote that detached from the archaebacterial branch after the first forking of the tree of life. In seeming conflict with this possibility, eukary¬ otes possess a few traits that appear to be derived from an ancient eubacterium. These anomalies have been variously attributed to convergent evolution, to hori¬ zontal gene transfer, and even, as proposed by the German investigator Wolfram Zillig,1 to a primeval fusion event between an archaebacterial and a eubacterial partner. The apparent mixed ancestry of eukaryotes has also been used in support of more radical proposals that picture the primitive eukaryotic ancestor as a cell ante¬ rior to prokaryotes and endowed with either DNA genes or even RNA genes.2 In this chapter, I shall assume that the eukaryotic branch issued from an ances¬ tral prokaryote, presumably of archaebacterial nature with some admixture of

THE AGE OF THE SINGLE CELL

138

eubacterial traits acquired one way or another. This hypothesis enjoys the greatest favor and agrees with most known facts. The problem facing us, therefore, is to retrace a pathway from the prokaryotic to the eukaryotic type of organization. This pathway must be mechanistically plausible, consistent with available evidence, and causally explainable by natural selection. At first sight, the task appears daunting. Place an average bacterium and even the most primitive of eukaryotes side by side and you observe differences of such magnitude as to make the conversion of one into the other almost unimaginable. Fortunately, one valuable piece of information is at hand. It is now established with a high degree of certainty that certain parts of eukaryotic cells, including mitochon¬ dria, chloroplasts, and, perhaps, peroxisomes (see chapter 17)—all three of which are membrane-bounded granules about the size of bacteria—are, in fact, the descendants of bacteria that were taken up by some eukaryote ancestor about 1.5 billion years ago and adopted permanently by this cell as endosymbionts (from Greek roots that add up in reverse to “living together inside”).3 Thus, the history of the eukaryotic cell may be divided into two separate eras: the pre-endosymbiont era—between 3.5 and 1.5 billion years ago—and the post-endosymbiont era—from 1.5 billion years ago to the present.4 Reconstructing the second era is no major problem. Cells take up bacteria all the time—our white blood cells do nothing else when we suffer an infection—and we know in detail the mechanisms involved in this uptake. To be sure, the bacteria taken up are usually killed and broken down by their captors, or, alternatively, attack their captors and kill them. A number of cases are known, however, in which this conflict leads to a stand-off that is solved by peaceful cohabitation. Thus, we have plenty of information to draw on for our reconstruction. The first era, which covers the transition from the ancestral prokaryote to a cell capable of capturing bacteria and adopting them as endosymbionts, is more myste¬ rious. But we have a few clues. First, if bacterial uptake took place in those remote times the way it does today, we can identify from present knowledge a number of properties that must have been gained in the course of the transition to provide the cells with the ability to capture. Also, a set of valuable clues are provided by sequencing results and other biochemical data that allow us to trace the prokaryotic ancestry of a number of eukaryotic structures. Finally, and most informative, a few primitive unicellular eukaryotes are known that date back to the pre-endosymbiont era and that may give us some idea of what eukaryotic life was like at that time. Let us take a look at such a “living fossil.”

GIARDIA, A LIVING FOSSIL The most ancient eukaryotes known are the diplomonads, a group that includes Giardia lamblia, a parasitic microorganism responsible for a number of severe

THE MAKING OF A EUKARYOTE

139

intestinal infections in humans and some animals.5 According to sequencing results, Giardia is at the end of a line that may have branched from the main eukaryotic trunk more than two billion years ago, that is, before oxygen started appearing in the Earth s atmosphere. The ^organism thus had an extremely long time to evolve and may bear little resemblance to the remote forebear it shares with other eukary¬ otes. It no doubt has changed, but probably not beyond recognition. It has so many features found also in more recent eukaryotes that we may safely attribute most of these features

there is always the odd possibility of evolutionary convergence, the

inevitable caveat—to a common pre-endosymbiont ancestor. Therefore, Giardia should provide us with useful information about this ancestor, leaving as gaps to be filled properties that were lost in evolution, for example, in the conversion of the organism from an autonomous to a parasitic way of life. Giardia is a single-celled, pear-shaped organism, about one-thousandth of an inch in size, which makes its volume more than 10,000 times bigger than that of an aveiage prokaryote. We are definitely dealing with a giant cell in comparison with bacteria. It is not encased within a solid wall, as are bacteria, but coated only by some fuzzy material without any rigidity. The cell nevertheless maintains its char¬ acteristic shape thanks to a number of inner props forming what is known as a cytoskeleton. One face bears a disk-shaped structure that serves as a sort of sucker enabling the cell to adhere to the intestinal wall, a feature no doubt acquired in the course of adaptation to its particular habitat. Giardia is highly mobile, propelled by four pairs of long, undulating flagella.6 These organelles are totally different from the motor appendages of bacteria bear¬ ing the same name; they are built of slender, hollow rods of protein nature, called microtubules, combined with many other proteins to form a long flexible shaft endowed with the ability to bend in a wavy movement, using energy provided by the splitting of ATP. Flagella share this basic structure with cilia, which are short, rapidly beating appendages usually present in large numbers on the surface of the cells they equip. The two kinds of motors never exist together on the same cell, as witnessed by the taxonomic distinction between flagellates and ciliates. Flagella and cilia are widely distributed throughout the eukaryotic world. We each owe our existence to the proper functioning of a flagellum that propelled a sperm cell emitted by our father into an egg cell presented by our mother. It is quite impossible that a structure as complex as a eukaryotic flagellum could have arisen independently twice by convergent evolution. Giardia thus tells us that the distant eukaryotic ancestor had already acquired all the main proteins that take part in the construction of flagella. It is particularly informative to watch Giardia feeding. It engulfs extracellular objects.7 It does what the ancestral cell that first came to harbor endosymbionts is known to have done. And it does so by a mechanism virtually identical to that whereby our own white blood cells and innumerable other eukaryotic cells eat bac¬ teria or other solid objects. Called phagocytosis (from the Greek roots for eating and cell), this process is so complex that we may assume, once again, that it existed

140

THE AGE OF THE SINGLE CELL

already in the pre-endosymbiont ancestor. We don’t have to search further to account for the manner in which endosymbionts entered their host cells. Just as we would have guessed solely from our knowledge of the present, these important guests were taken up by a typical phagocyte endowed with the basic attributes of similar extant cells. This is an invaluable piece of information. Lynn Margulis, a University of Massachusetts biologist known for advocating the endosymbiont the¬ ory at a time when there was little evidence to support it, has postulated aggressive invasion by a “fierce predator” to account for endosymbiont entry.8 In my opinion, Giardia makes this hypothesis unlikely. Uptake by phagocytosis seems a more probable explanation, supported by all that we know. Let us look at Giardia—or one of its free-living cousins—on the hunting trail. See it accidentally brushing against a bacterium or, perhaps, moving toward it in a seemingly purposeful manner, attracted by a chemical signal, as would our white blood cells in a similar situation. Whether chance or chemistry caused the encounter, the consequence is that the contacted bacterium remains stuck to the cell’s surface, like a fly to flypaper. Triggered by this contact, the organism goes into action, slowly sucking in its hapless victim, which progressively vanishes from our sight. Soon, no trace of this dramatic gobbling remains on the cell’s surface, which has regained its smooth, unruffled appearance. Not all bacteria are caught in this way. This is because sticking requires the pres¬ ence of receptors on the captor’s surface that recognize certain components on the bacterial cell wall—another case of a complementary lock-and-key arrangement. If the lock or the key is missing, the bacterium simply bumps away after the collision and escapes. Some bacteria actually take advantage of such a defect to elude cap¬ ture, a fact that has played a tremendous role in the history of science. It so happens that infectious pneumococci, the redoubtable agents of bacterial pneumonia, differ from their harmless relatives by lacking the gene coding for a cell-wall component that is specifically recognized by our white blood cells. They can’t make the lock, and so the blood cells’ key does not find a fit. In 1928, Fred Griffith, a medical offi¬ cer in the British Ministry of Health, closely followed by the American Martin Dawson working at the Rockefeller Institute for Medical Research (now the Rocke¬ feller University) in New York, found that the genetic ability to make the cell-wall molecule that renders microbes catchable and thereby harmless could be transferred from dead, noninfectious bacteria to live, pathogenic ones. Sixteen years later, three Rockefeller scientists, Oswald Avery, Colin MacLeod, and Maclyn McCarty, announced to an initially incredulous world that they had purified the “transforming factor” and identified it beyond reasonable doubt as a DNA molecule.9 This historic experiment established for the first time that genes are made of DNA, not of protein as many had believed. It set off one of the most epic scientific races, which was won in 1953 by Watson and Crick when they deciphered the double helix. When these investigations were performed, the mechanism of phagocytosis was not yet known. Today, thanks to electron microscopy and other sophisticated tech-

THE MAKING OF A EUKARYOTE

141

mques, we understand the phenomenon in detail. The caught bacterium does not enter the cell through a hole, as one might suspect. It is progressively enveloped by a deepening infolding, or invagination, of the cell membrane, which remains intact. When envelopment is completed, the invagination snaps off from the inner surface without leaving any trace and turns into a closed intracellular sac, or vesicle, with the engulfed bacterium inside, entirely surrounded by the piece of membrane abstracted from the cell surface in the course of engulfment. The fluidity, flexibil¬ ity, and self-sealing properties of lipid bilayers explain how such a phenomenon is physically possible. What drives the uptake? We don’t know about Giardia, which has not been stud¬ ied in this respect, but we know of at least two mechanisms involved in other cells. One, external, relies on progressive “zipping”—better said, “Velcro-ing,” if such a word existed, since we are dealing with surfaces—between the phagocyte’s recep¬ tors and their complementary partners on the bacterial surface. The other mecha¬ nism is activated from inside the cell by a device that draws in membrane patches whose receptors are occupied. This mechanism allows the uptake of fluid droplets, or pinocytosis (Greek for “cell drinking”), triggered by the binding of soluble mole¬ cules to surface receptors. The general term endocytosis covers all forms of mem¬ brane-dependent engulfment, including phagocytosis and pinocytosis. The intracel¬ lular vesicles formed by endocytosis are called endosomes. Subsequent events bring the catch into another kind of membrane-bounded intracellular vesicles called lysosomes (“digestive bodies,” in Greek), in which engulfed materials are exposed to acid and to digestive enzymes, thereby suffering the same fate as food in our stomach. The acid is secreted into the lysosomes by a proton pump present in the lysosomal membrane, whereas the digestive enzymes are conveyed to lysosomes from another set of intracellular vesicles named the endoplasmic reticulum, ER for short. As digestion proceeds within lysosomes, the small nutrient molecules arising from this process pass through the lysosomal membrane into the cell proper, to participate in metabolism. In the end, the lyso¬ somal content of enzymes and undigested residues is often unloaded into the extra¬ cellular medium by a process graphically called cellular defecation, or exocytosis in general, essentially the reverse of endocytosis. In this process, a vesicle adds its membrane by fusion to the cell membrane and discharges its contents outside. The digestive enzymes destined to be conveyed to the lysosomes are simultane¬ ously (cotranslationally) made and translocated into the interior of ER vesicles by ribosomes closely apposed to the vesicles’ membranes. These parts of the ER, which are studded with ribosomes, are called “rough” because of the rugged appearance of the membranes in cross section. From these rough parts, the newly made enzymes are transferred to smooth (ribdsome-less) parts of the ER and, sub¬ sequently, to a complex of membranous sacs known as the Golgi apparatus, system, or complex—Golgi for short—from the name of the Italian neuroanatomist Camillo Golgi, winner, with his Spanish colleague Santiago Ramon y Cajal, of the 1906

THE AGE OF THE SINGLE CELL

142

Nobel Prize in medicine. In the course of the passage of the enzymes through the smooth ER and the Golgi, their molecules undergo a number of chemical modifica¬ tions generally referred to as processing or maturation. Together, endosomes, lysosomes, rough and smooth ER, and Golgi form a com¬ plex intracellular network of sacs and vesicles, sometimes called the cytomembrane system. Consisting of up to thousands of distinct membrane-bounded compart¬ ments, this system governs a number of important cell functions grouped under the general labels of bulk import and export. Import takes place by endocytosis and usually leads to digestion in lysosomes, occasionally to storage or transcellular pas¬ sage of the material taken up. Export starts in the rough ER, continues by way of the smooth ER, Golgi, and lysosomes, and ends in defecation by exocytosis. Export also follows an alternative pathway, much more important in most cells, that bypasses lysosomes and leads directly from the Golgi to exocytic discharge. This is the major pathway of secretion. Traffic of the imported and exported materials through the many compartments of the cytomembrane system is mediated by permanent or, more often, transient connections between compartments, and directed by external railings belonging to the cytoskeleton and by internal receptors. The energy required to drive these movements comes from the splitting of ATP, with the help of special cytoskeletalmotor systems. The cytomembrane system is a unique characteristic of all eukaryotic cells. Giardia tells us that this system was already laid out two billion years ago. Even the Golgi-dependent secretion machinery had been developed by that time.10 There are no mitochondria, chloroplasts, or other possible descendants of engulfed bacteria in Giardia. Although such organelles may well have been jetti¬ soned in the course of evolution, it is of interest that the next most ancient known eukaryotes, the microsporidia, show a similar lack of endosymbiont-derived organelles.11 These facts, together with other pieces of evidence, strongly support the view that these very ancient organisms descend from lines that branched from the eukaryotic trunk before endosymbionts were adopted. This lineage makes the organisms the closest extant relatives—although still immensely remote—of the pre-endosymbiont eukaryote. Metabolically, Giardia is an obligatory anaerobe adapted to an oxygen-free environment. This adaptation could conceivably go back to when Giardia's distant forebear detached from the eukaryotic line, since oxygen had not yet started to appear in the Earth’s atmosphere at that time. If such is the case, Giardia's entire ancestral lineage survived through the oxygen crisis and continued evolving for more than one billion years, somehow protected from oxygen until the time it found an appropriate oxygen-free environment in the gut of some animal. It is equally possible—more probable?—that the ancestral cells became adapted to oxy¬ gen and subsequently lost this property when they adjusted to anaerobic parasitic life. It is, however, noteworthy that the main electron-transfer chains of eukaryotes

THE MAKING OF A EUKARYOTE

143

belong to endosymbiont-derived organelles, not to the cell membrane or the cytomembrane network. Barring an evolutionary loss, this fact suggests that the piimitive eukaryote, which was probably anaerobic, like all organisms that existed more than two billion years ago, may have lacked membrane-bound respiratory chains and that its descendants managed to weather the oxygen crisis without the help of such chains until endosymbionts came to the rescue. The genetic organization of Giardia appears to be of the “classical” type, as far as is known. Two points are of interest. First, the organism’s ribosomes are more similar, in terms of certain molecular characteristics, to prokaryotic than to eukary¬ otic ribosomes. This similarity is in keeping with an early branching from the eukaryotic line, before the present kind of eukaryotic ribosomes had evolved. Next, and particularly interesting, Giardia has two nuclei of equal size. But before we look at this intriguing duplication, let us consider the nuclei themselves, the hall¬ mark of eukaryotic cells.

THE EUKARYOTIC NUCLEUS Giardia's nuclei have all the main features of eukaryotic nuclei. So named because it sits like a kernel (nucleus in Latin, karyon in Greek) in the middle of the cell, the nucleus is a voluminous, roughly spherical body entirely surrounded by a doublemembranous envelope structurally and functionally related to the ER (the outer membrane is studded with ribosomes). The inner face of this envelope is bolstered by a sturdy lining made of tightly knit protein fibers. A large number of reinforced openings, or pores, inserted like portholes through the nuclear envelope, serve as regulated passageways between the nucleus and the rest of the cell, or cytoplasm. The main residents of the nucleus are the chromosomes (Greek “for colored bodies”), so named not because they are colored but because early microscopists saw them as intensely stained bodies in cell preparations treated with certain dyes. Eukaryotic chromosomes are majestic edifices compared with their prokaryotic counterparts, which are little more than a circular stretch of naked DNA. In con¬ trast, the eukaryotic chromosome is a highly structured entity. Imagine a miniature maypole wreathed by loops of a spirally wound garland of beads. The pole is the inner skeleton of the chromosome, constructed of protein. In the garland, the string is made of DNA and the beads consist of small protein spools around which the DNA string makes a couple of turns before moving on to the next spool. This beaded string is twisted into a thick thread, somewhat in the form of a telephone cord. The thread itself is divided into a series of ample loops anchored to helically disposed attachment points around the central skeleton. As a rule, some of the loops are uncoiled, others are packed into tight balls. When a cell starts dividing, all the uncoiled loops also become packed into balls and the chromosome assumes the

THE AGE OF THE SINGLE CELL

144

shape of a thick, knobby rod. It is in such dividing cells that chromosomes were first observed as stained rods. In nondividing cells, the basic chromosome structure is hidden by the inextricable tangle of uncoiled DNA stretches. And what a tangle it is! Imagine some two miles of a very thin string looped around a two-foot rod. This is how your average chromosome would look magnified 100,000-fold. Existence of a nucleus entails a number of fundamental consequences that make the eukaryotic organization totally different from the prokaryotic organization from which it originates. First, the nuclear envelope separates the cell into two distinct compartments that communicate with each other only by the nuclear pores. This kind of compartmentalization differs from that created by the cytomembrane net¬ work, which, with its multiple interconnected cavities lined by chemical-processing machineries and transport systems, forms a sort of halfway house between the inside of the cell and the outside world. The nuclear envelope divides actual metab¬ olism. The principle of this division is simple. Keep in the nucleus only those func¬ tions that have a close link with DNA and leave all the rest in the cytoplasm. To maintain proper connections between the two, equip the pores with specific sys¬ tems that mediate the passage of all the substances that need to go in or out, under strict control of their molecular identity. Two functions obligatorily situated inside the nucleus are DNA replication and DNA transcription. Because of the complications resulting from chromosome struc¬ ture, DNA replication requires a complex variety of disentangling systems to make the DNA accessible to the replicating enzymes. For this reason, the process is at least twenty times slower in eukaryotes than in prokaryotes. This drawback is offset by having the DNA distributed over several chromosomes—there are four in Giardia, forty-six in human cells—and by having multiple replication sites on each chromosome. In prokaryotes, there is a single replication site (anchored to the cell membrane) through which the entire chromosome is reeled for replication. In eukaryotes, there are a large number of such sites, so that the DNA is replicated simultaneously in many short stretches, which are subsequently joined together. Thanks to this arrangement, the entire eukaryotic genome (about six feet of DNA in a human cell) can be replicated in about one hour, little more than twice the time needed for the replication of a prokaryotic genome only one-twentieth of an inch long. When DNA replication proceeds in a eukaryotic nucleus, all the proteins needed for the construction of chromosomes are imported into the nucleus from the cytoplasm and assemble spontaneously with the newly formed DNA to form a sec¬ ond set of fully formed chromosomes, each joined by a bridge to its pre-existing sister, like Siamese twins. Intranuclear DNA transcription runs into the same kind of structural problems as DNA replication, with the additional complication that the synthesized RNA prod¬ ucts are exported out of the nucleus. Only mature RNAs are sent out to the cyto¬ plasm. Splitting, trimming, splicing, and other RNA rearrangements are all carried out in the nucleus. A special intranuclear organelle called the nucleolus harbors the synthesis and maturation of ribosomal RNAs (the RNAs that combine with a set of

THE MAKING OF A EUKARYOTE

145

proteins to form the ribosomes), which constitute by far the largest part of the RNA output of the nucleus at any given time. Other complex systems situated in the nucleus ensure the splicing of messenger RNAs, a major function in higher eukary¬ otes, but perhaps not in Giardia, which, with no split genes detected so far, may have no need for RNA splicing. Mature RNAs do not move out of the nucleus on their own, but in the company of special RNA-binding proteins, which are admitted into the nucleus unaccompanied and return to the cytoplasm with their quarry. An important consequence of the kind of segregation just described is that the translation of genetic messages is topologically separated from their transcription. This is not so in prokaryotes, where one can often see ribosomes busily making proteins on messenger-RNA stretches that are still in the process of being tran¬ scribed from DNA. In eukaryotes, the ribosomes are all in the cytoplasm, where they function with messenger-RNA molecules that have been synthesized and prop¬ erly processed in the nucleus, allowed to pass through the nuclear pores, and offered to the ribosomes in intact and accessible form. Gene expression can thus be regulated at multiple points, both in the nucleus and in the cytoplasm. The existence of the eukaryotic nucleus creates special problems for cell divi¬ sion. In prokaryotes, cells divide after chromosome duplication by a simple con¬ striction, or furrow, of the cell membrane (and of the outer wall), which takes place in such a way that each daughter cell inherits one of the duplicated chromosomes with its half of the original cell membrane. In eukaryotic cells, the cytoplasm divides by a roughly similar mechanism, but not before the nucleus has itself undergone duplication. Nuclear division is a dramatic phenomenon, one of the few cellular events that can be observed in some detail with an ordinary light microscope. It has fascinated generations of biologists. In a “nutshell,” the chromosomes are first duplicated and compacted. This is when they become visible as rods or filaments, which explains the name of mitosis given to nuclear division (mitos means thread in Greek). Next, the nuclear envelope is dismantled and replaced by the spindle, a complex rigging constructed from microtubules, the same structures that form the flagellar shaft. The duplicated chromosomes then congregate on the equatorial plane that divides the spindle in two halves. The rigging now comes into action, forcibly wrenching paired chromosomes away from each other and pulling one member of each pair toward one of the two poles of the spindle. Here we see the advantage of the Siamese-twins structure of duplicated chromosomes. It allows the paired chromo¬ somes to align in a proper orientation and the rigging to work in such a way that a full identical set of chromosomes is assembled at each pole of the spindle. When this assembly process is completed, the spindle is dismantled and brand-new nuclear envelopes form around each chromosome set. Giardia has characteristic nuclei that divide by typical mitotic division (except that, as in a number of primitive protists, the nuclear envelope does not break down).12 We may take it, therefore, that the primitive eukaryote already possessed all the relevant structures and properties that go into the making and division of

146

THE AGE OF THE SINGLE CELL

eukaryotic nuclei. But why should Giardia have two nuclei instead of the single nucleus commonly found in eukaryotic cells? And, as a corollary to this question, could the primitive eukaryote also have possessed two nuclei at a certain stage of evolution? We don’t know the answer to these questions. But we can think of one with such tremendous implications that it deserves a separate treatment. It has to do with the origin of the most powerful force in nature: sex. I shall consider it in the next chapter.

Chapter 16

The Primitive Phagocyte >>

The picture is clear. When Giardia’s lineage branched from what was to

become the main eukaryotic trunk, probably more than two billion years ago, almost all the key features of eukaryotic cells, with the exception of endosymbiontderived organelles, had already emerged. The crucial prokaryote-eukaryote transi¬ tion occurred some time during the 1.0 to 1.5 billion years following the primeval forking that led to the eukaryotic branch. During that time, a simple prokaryote developed into a primitive phagocyte, a large nucleated cell capable of capturing food and digesting it intracellularly. What pathway did this momentous transforma¬ tion follow? And, especially, why was this road actually taken in reality? Extant organisms offer a number of valuable clues to the first question, but we have only educated guesses to help us answer the second. Remember the rule: fore¬ sight excluded. There was no goal, no eukaryotic ideal beckoning from the distant future, inviting evolving cells to overcome hurdles and vanquish difficulties. Every step of this extraordinary voyage was taken in its own present context, the conse¬ quence of some chance mutation that happened to confer an immediate benefit favoring the survival and proliferation of the affected cell there and then. What hid¬ den selective forces cut open this trail, step by step, over an immensely long period of time, to produce what was probably the most epoch-making innovation in the history of life? This question will be with us as we try to retrace the main steps of the voyage. From what we have seen of Giardia, there are really two major developments to be accounted for within the context of an enlarging cell: cytomembranes and cytoskeletal elements, with a fenced-off nucleus arising through a special combina¬ tion of the two. We have no clues to the origin of the cytoskeleton, which may be a true innovation. But we know the origin of eukaryotic cytomembranes. According to all available evidence, they come from the ancestral prokaryotic cell membrane.

148

THE AGE OF THE SINGLE CELL

SPREAD OF A NETWORK A banal accident, which turned out to have long-term effects of enormous impact, may have initiated the whole chain of events. An ancient heterotrophic prokaryote lost the capacity to build a cell wall. This defect most often weakens survival ability but is not invariably lethal. Naturally wall-less bacteria are known, including highly thermophilic ones. In this particular case, the circumstances surrounding the acci¬ dent were such that the victim not only survived but benefited from its infirmity. Perhaps the maimed organism was a resident of one of those multilayered bacterial colonies that were beginning to flourish at that time and have left their traces in the form of stromatolites. Living in the midst of bacterial mats, our denuded remote forebear was sheltered from many hazards and suffered little from its nakedness. It could continue to proliferate at the expense of its neighbors and produce similarly naked progeny. According to fossil evidence, stromatolite colonies have persisted with little apparent change from the earliest days of life to the present. If, as is pos¬ sible, the prokaryote-eukaryote transition required a stable food-supplying environ¬ ment for a very long time, colonies of this kind could have satisfied such a require¬ ment. Another event that may have happened very early is the acquisition of mem¬ brane lipids of the ester type. All eukaryotes have ester lipids. This is one of the eukaryotic features that do not fit with an archaebacterial origin. Ester lipids are characteristic of eubacteria, whereas all known archaebacteria have ether lipids. There are many possible explanations for this discrepancy, and I shall not go into them. Let us simply record that our putative wall-less ancestor probably had ester lipids, which means that it may have lived in a milder environment than did the thermophilic bacteria from which it presumably originated. In addition, the posses¬ sion of ester lipids, through the increased membrane fluidity it conferred, could have been an important factor in the process whereby the loss of a cell wall turned into a benefit for heterotrophic life. In order to appreciate this benefit, let us visualize our stripped ancestral cell. It is a shapeless, flattened blob that feeds on the remains of dead bacteria, which it digests extracellularly by means of secreted enzymes, as do all heterotrophic prokaryotes. Here is where nakedness becomes an advantage. There is no straightjacket around the cell, no barrier between cell and food supply. Helped by the flexi¬ bility of its membrane, the cell can stick intimately to the bodies on which it feeds, mold itself to their contours, and even wrap itself completely around them, helped in all these movements by surface receptors, or binding sites, capable of hooking on to certain surface components of the bacterial bodies. Thanks to these intimate con¬ tacts, the digestive enzymes discharged through the cell membrane remain trapped between cell and prey and can act optimally. In turn, the small nutrient molecules produced by digestion can enter the cell readily across the cell membrane, again without loss or delay. Our naked heterotroph is a magnificent feeder, a born winner

THE PRIMITIVE PHAGOCYTE

149

in the struggle for food, as long as its environment is protective enough to offset the lack of a cell wall. Second advantage: Our hero can grow bigger. The size of a cell is limited by the surface area it has available for exchanges with the outside—of nutrients inward and of waste products outward. A spherical cell surrounded by a smooth membrane cannot grow beyond a limit size because the volume increases with the third power of the radius, and the surface area only with its second power. To grow bigger, the cell must either change its shape—to a rod or filament, for example—so that it has more surface available for a given volume, or expand its membrane by infoldings (invaginations) or outfoldings (evaginations). Our naked ancestor was a champion contortionist in this respect. It could grow to any size by pleating its surface.1 But why should it do this? More efficient feeding is a likely answer. The more jagged a coastline, the more sheltered the inner coves within which two partners— digestive enzymes and food in the present case—can meet without disturbance. And so, natural selection would favor a larger cell with a more irregular contour. The final outcome is predictable by anybody acquainted with the self-sealing habits of lipid bilayers. As invaginations deepen, the gullets leading into them narrow until—click—there is no gullet anymore. The invagination has snapped off from the surface to form a closed vesicle inside the cell, while the scar left in the cell membrane by the amputation heals simultaneously by self-sealing. A small bubble blown into the surface of a larger bubble suddenly cuts its moorings and floats ghostlike inside—a trick some soap-bubble experts actually can produce. Inside the small inner bubble, food and digestive enzymes are now completely segregated together. From being extracellular, digestion has become intracellular.2 The trick, when it happened naturally, was not a trivial event. It inaugurated a crucial development in cellular evolution: the phagocytic way of life. For the first time, heterotrophs had a stomach of their own. No longer compelled to carve a stomach out of their surroundings and to reside within it, they could now afford to wander around and to survive by capturing food. This was a gigantic step on the way to cellular emancipation. From a captive held in the golden prison of a nutri¬ tive shell, like a maggot in a chunk of cheese, the cell had turned into a mighty hunter that could invade the world in search of prey. The first cell stomach was a compound of many things. Arising from an engulfment phenomenon, it acted as a storage place for engulfed food. At the same time, the primeval stomach received digestive enzymes from ribosomes bound to its membrane. These ribosomes simply continued the job they performed on the cell membrane, with the difference that the enzymes were now collected inside the stomach, instead of being discharged into the outside medium, and acted directly on the collected food. By virtue of its origin, the stomach’s membrane possessed all the transport systems present on the cell surface. These included a proton pump originally directed outward and now toward the inside of the stomach, making it acidic and thereby optimally suited to the requirements of the digestive enzymes. Like our stomach, the cell’s stomach needs acid for the best performance of its

150

THE AGE OF THE SINGLE CELL

enzymes. Other transport systems, which once transferred into the cell’s interior the small nutrient molecules produced by extracellular digestion, now cleared the stomach of digestion products in the same way. Yet others, acting in the reverse direction, discharged into the stomach the waste products they previously excreted into the outside medium. Also present, protruding on the inner face of the stom¬ ach’s membrane, were the cell’s surface receptors, including those that had caused the cell to stick to some materials and engulf them. Finally, it often happened that by a reversal of the manner in which it first arose, the stomach joined back with the cell membrane, thus restoring to this membrane the patch that had been removed from it by the original internalization phenomenon, and at the same time unloading into the outside medium the stomach’s contents of undigested food, waste products, and enzymes. Put all this together and you find that the first cell stomach combined functions that, in higher organisms, would be described as ingestion, secretion, digestion, absorption, excretion, and defecation. Thanks to our foray into the future, we iden¬ tify in this primeval stomach properties typical of endosomes, rough ER vesicles, and lysosomes, all in one, and we recognize as endocytosis and exocytosis the two phenomena that link the stomach reversibly to the cell membrane. Further evolu¬ tion can be summed up as a progressive segregation of the various functions carried out by the primeval stomach to separate parts of an increasingly complex network of intracellular vesicles, all derived from the ancestral cell membrane. In an analo¬ gous way, but on an entirely different scale, the same functions are distributed from mouth to anus all along our own digestive tract, whereas they are carried out by a single cavity in a primitive organism such as a jellyfish. The first functions to become dissociated were food collection and enzyme stor¬ age. This was accomplished by further membrane internalization and by migration of the ribosomes previously associated with the cell membrane to a new set of intracellular pouches. Ancestral to the rough ER, these pouches turned into recepta¬ cles of newly synthesized digestive enzymes, which were no longer discharged directly by the membrane-bound ribosomes into the outside medium or into invagi¬ nations and vesicles derived from the cell membrane. Because of this migration phenomenon, capture of food was now accomplished by ribosome-free membrane patches, and the resulting vesicles served in the temporary storage of engulfed food but not in its digestion. The initial phenomenon of haphazard membrane infolding thereby turned into what we now know as endocytosis, and the internalized vesicles became endosomes. Receptors on the cell-membrane surface made it possible for the cells to be selective and to choose their “menu” from the materials present in the surrounding medium. The stomach proper, or lysosome, was created as a separate, acidified compart¬ ment lying between the enzyme-containing, rough-surfaced pouches and the foodcontaining endosomes, and connected to these two sites by a variant of the bubble trick called vesicular transport. In this process, vesicles bud off from one site, car-

THE PRIMITIVE PHAGOCYTE

151

rying with them material stored in that site, and dock at another site where they deliver their contents. In terms of the bubble analogy, a small bubble detaches from a larger one, as may happen to a soap bubble caught in an air current. The drifting, small soap bubble next bumps into another, large soap bubble and merges with it. As a result, some bubble material and a small volume of air are transferred from one soap bubble to another. In vesicular transport, membrane fabric is likewise transferred from one closed sac to another, though not with a small amount of air but with important enclosed materials. In this way, food transferred from endosomes, and enzymes transferred from rough-surfaced pouches, were led to con¬ verge into lysosomes, where the one could now be attacked by the others. Lysosomes did not swell indefinitely as a result of this dual transfer. Small mole¬ cules arising through digestion were transferred to the cell interior by transport sys¬ tems (inherited from the cell membrane) situated in the lysosomal membrane, whereas residues left at the end of digestion were unloaded from the lysosomes into the outside medium by exocytosis. As to the excess membrane material, part was removed from the lysosomal membrane by vesicles shuttling back empty to their sites of origin, and the rest was added to the cell membrane by fusion in the final phenomenon of exocytic discharge. Thanks to this continual recycling, membrane material remained stably distributed among the cell membrane and the different parts of the intracellular membrane system, in spite of the intense traffic taking place among them. Going back to our comparison with the animal digestive tract, we have reached a stage where there is a mouth (endocytosis), a digestive cavity (the lysosome), and an anus (exocytosis), with an attached digestive gland pouring in digestive juices in the manner of our pancreas (rough ER). An important difference, besides the enor¬ mous difference in scale and, thus, in the nature of the structures involved, is that the parts are not linked by continuous channels controlled by valves, but by inter¬ mittent connections established by vesicular transport. In either case, the inside of the system is kept separated at all times from the rest of the body, except for the transfer of selected materials across the tract’s lining, mediated by specific machineries. During the subsequent evolution of this primitive intracellular digestive tract, additional intermediate stations were inserted into the main traffic lines to serve for temporary storage and specific chemical processing of the passing materials, or for their sorting and selective rerouting by means of special receptors. In this way, the smooth parts of the ER and the different components of the Golgi complex became intercalated between the rough ER and the lysosomes. The endosomes, on the other side, became subdivided into several sections, which allowed some of the materials taken up to be saved from lysosomal digestion and diverted to other directions inside or outside the cell. A major diversion came to be inserted between the exit from the Golgi and the lysosomes, so that materials traveling along the ER-Golgi pathway were conveyed

152

THE AGE OF THE SINGLE CELL

directly to the cell surface for discharge into the outside medium, without passing through the lysosomes. Eventually, this line turned into the major pathway of secre¬ tion, the process whereby cells discharge around them the components of extra¬ cellular structures and a variety of complex materials, such as enzymes, hormones, and other active agents, that they manufacture for export. These substances all con¬ sist ol proteins that are made in the rough ER and are further trimmed and fitted with a variety of carbohydrate, lipid, and other components as they pass through the smooth ER and the Golgi. This new line bypassed the lysosomes, thus avoiding damage to the channeled materials. The original line leading to lysosomes was maintained, but under the control of receptors that admitted only molecules bearing a specific chemical tag common to all the digestive enzymes destined for lyso¬ somes. Swept up in the general process of membrane internalization was a special patch of prokaryotic cell membrane to which the chromosome was hooked by the system serving in DNA replication. Such attachment is a common feature of all prokary¬ otes. Internalization of this specialized membrane part moved the chromosome and its replicating system to the cell interior, where they became progressively sur¬ rounded by a sealed double-membranous envelope derived from the cell membrane and structurally and functionally connected with the rest of the intracellular mem¬ brane network. This membrane migration was a development of cardinal impor¬ tance. It initiated the birth of the nucleus, the cell part to which eukaryotes owe their name. My historical reconstitution of the development of the eukaryotic cytomembrane network is hypothetical, since descendants of intermediate forms have not been uncovered. However, the model is supported by a great deal of information written into the structures and functions of existing molecules. Scattered throughout the intracellular membrane network of eukaryotes are unmistakable molecular relatives of systems associated with the cell membrane of prokaryotes: characteristic trans¬ port systems in one part of the network, translocating ribosomes in another, lipidsynthesizing enzyme complexes in another, translocating carbohydrate-assembly systems in another, traffic-directing receptors in another, chromosome attachments in yet another, and so on. Formation of the eukaryotic membrane network by inter¬ nalization and differentiation of an ancestral prokaryotic cell membrane is highly probable. Only the details are missing. Any model of evolution must demonstrate that almost every proposed step pro¬ vided a selective advantage. I have taken as the main driving force the progressive conquest of greater heterotrophic autonomy through the enhanced ability to find, take up, and utilize food, which, for a heterotroph, is the key condition of survival and reproductive success. This explanation is plausible and makes sense in the con¬ text of present knowledge. Furthermore, it allows for the progressive unfolding of an evolutionary process that comprised a very large number of successive steps and stretched out over a very long duration. Each little step of the proposed scenario may be seen as associated with a slight improvement in phagocytic efficiency.

THE PRIMITIVE PHAGOCYTE

153

THE INDISPENSABLE PROPS AND MACHINES Membrane expansion and internalization alone could not have brought about the de¬ velopments I have described. What the cells needed in addition were internal props—a cytoskeleton—that would save their growing bulk from collapse without, however, impairing their protean ability to change shape. In addition, the cells needed motor systems to perform the work involved in uptake, transport, and dis¬ charge of materials through the different cavities of the growing membrane network. An astonishing number of complex molecular systems emerged to satisfy these re¬ quirements. It is remarkable that related assemblages have so far not been detected in prokaryotes. Bacterial supporting structures and flagella are entirely different from their eukaryotic counterparts. Unlike the cytomembrane network, which clearly originates from the ancestral cell membrane, eukaryotic cytoskeletal and motor sys¬ tems seem to be true innovations made during the prokaryote-eukaryote transition. This makes their emergence particularly critical to the transition. Unfortunately, no clue to the origin of these elements has yet been uncovered. They are either absent or present in their fully sophisticated form. There is no sign of the many intermediate forms that must have marked their development. Many intracellular and extracellular structures are built of long, threadlike mole¬ cules, either proteins or carbohydrate polymers, that are intertwined into a variety of fibers, bundles, webs, sheets, plates, baskets, and other three-dimensional arrangements. As a rule, these structures are stable and static. In multicellular organisms, they give many cell types their specific shape or provide external frame¬ works for cells to assemble into characteristic tissues, such as those that make up skin, bones, joints, mucosae, viscera, and so on. Structures of this kind would have been of little help to our fledgling phagocyte. They would simply have replaced an external straight]acket by an internal one. What evolution came up with, not once but at least three times, are Lego-like pro¬ tein molecules capable of assembling reversibly into a rigid arrangement. Two of these molecules, actin and tubulin, are constructed according to similar principles. Imagine a set of identical building blocks that can join together by some sort of complementary peg-and-hole devices, like the pieces ol construction toys. Each block has a peg at one end and a hole at the other, so that blocks can be linked together indefinitely in a linear fashion to form a rod or thread. In addition, each block also has a peg on one side and a hole on the other, allowing the threads to join laterally. In actin, this lateral link is such that two threads wind into a double-helical fiber. The link is such in tubulin that thirteen threads join spirally into a hollow, cylindrical tube, or microtubule. Actin fibers and microtubules have in common that they can be dismantled and reassembled into different configurations with the help of ATP. They thus serve to shore up cells into a variety of temporary shapes, which sometimes even initiate

154

THE AGE OF THE SINGLE CELL

movements by their changes. Rigidity is allied with plasticity. Actin fibers and microtubules are often connected with special ATP-splitting proteins that change shape when they split ATP and thereby act as converters of chemical energy into mechanical work. Both structures also participate, together with their associated molecular motors, in the construction of stable edifices of great complexity, which underlie the most elaborate forms of eukaryotic motility. We have already encountered microtubules as important cytoskeletal elements in Giardia. This organism holds two magnificent specimens of the extraordinarily var¬ ied structures that can be built with microtubules and their associated motors. One, transient, is the mitotic spindle, which is assembled at each cell division and dis¬ mantled again afterwards. The other, stable, is the flagellum, a slender, cylindrical structure built from nine parallel pairs of partly fused microtubules surrounding an axial shaft made of two microtubules (the 9 + 2 structure, also characteristic of cilia). Some five hundred additional proteins complete the assemblage. Among them is a special ATP-splitting protein, named dynein, that has the remarkable property of bending forcibly when it splits ATP The mechanical and the chemical process are obligatorily coupled. One cannot happen without the other. Thus, if the two ends of dynein are attached to separate structures, the molecule causes the structures to move closer to one another, using the energy released by the splitting of ATP to overcome an opposing resistance. This phenomenon is responsible for the wavy movement whereby flagella propel cells forward. The presence in the most ancient known eukaryote of structures that are among the most elaborate molecular assemblages found in the whole living world is an impressive fact. It suggests strongly that the development of such structures played an essential role in the prokaryote-eukaryote transition, perhaps to the point of set¬ ting the pace of this transition, as we are obviously dealing with an extremely long succession of evolutionary steps. To my knowledge, actin has not been detected in Giardia, or, for that matter, searched for in this organism. Thus, we don’t know whether actin is as ancient as tubulin. This seems likely, however, since actin is found in a variety of protists, as well as in all higher eukaryotes, often in the form of variously disposed bundles aligned against the cell membrane or in the form of cables stretched across the cell like telephone wires. Bundles of actin fibers are often joined end to end by axial shafts made of molecules of an ATP-splitting motor protein called myosin. When supplied with ATP and activated by calcium ions, the myosin shaft acts like a ratchet pulling the two actin bundles toward each other. Depending on what cell parts the actin filaments are attached to, this movement may cause all kinds of intracellular displacements and cellular deformations, including a sort of crawling known as amoeboid movement from the name of a protist, the amoeba, that typi¬ cally moves in this way. The most elaborate actin-myosin arrangements exist in animal muscle cells. They consist of parallel arrays of interdigitating actin filaments and myosin fibers,

THE PRIMITIVE PHAGOCYTE

155

held together by a number of additional proteins to form the muscle fibrils. These beautiful structures have given electron microscopists some of their most pleasur¬ able aesthetic experiences, -rivaled only by those provided by the contemplation of flagella, cilia, and other microtubule assemblages. A third kind of protein building block capable of assembling with the help of energy provided by the splitting of ATP is clathrin, which has a peculiar threelegged shape that allows assembly of many molecules into two-dimensional hexag¬ onal meshes of variable curvature, resembling the famed geodesic dome built by the American architect Buckminster Fuller. This structure plays a key role in recep¬ tor-mediated endocytosis and in some forms of vesicular transport. When receptors become occupied on the outer face of a cell-membrane patch, they undergo a con¬ formational change that causes clathrin molecules to be recruited on the inner face of the patch and to assemble against it into a closely adhering mesh, with the expenditure of ATP. The mesh progressively rearranges into domes or baskets of increasing curvature, drawing with it the adhering membrane patch with its attached prey and finally cutting it off from the rest of the membrane in the form of a closed, membrane-bounded vesicle containing the catch and surrounded itself by a clathrin trellis (which soon unravels). Although not yet detected in Giardia, clathrin could very well also have a very ancient history, in view of its many associ¬ ations with membrane movements. Another ancient cytoskeletal structure of paramount importance is represented by the internal supporting shell and associated pores of the nuclear envelope. Made of many different proteins, this complex structure dismantles, together with its dou¬ ble-membranous covering, with every mitotic division, and it reforms sponta¬ neously around each set of daughter chromosomes at the end of mitosis. This reconstruction phenomenon is one of the most remarkable known instances of the spontaneous assembly of a complex structure. The recipe is astonishingly simple. Take some juice from dividing cells, throw in any odd piece of naked DNA—even DNA that has never been near a eukaryotic cell—add a little ATP, and, lo and behold, in a matter of two to three hours, a perfectly respectable envelope assem¬ bles around the DNA, complete with double membrane, inner lining, and pores. Inside this mininucleus, the DNA even has formed a beaded string coiled into a miniature chromosome. In this process, hundreds of distinct pieces that were scat¬ tered around in the cell extract come together in seemingly miraculous fashion, summoned by no more than the added DNA and supplied with energy by ATP. We are not far from Hoyle’s Boeing 747 arising ready to fly from a tornado-swept junk¬ yard, except for a fundamental difference: There is information in all the pieces. They are not junk but pieces of a jigsaw puzzle, shaped to occupy a specific loca¬ tion in the overall picture. In contrast with the puzzle pieces, however, which are cut from a pre-existing picture, the nuclear building blocks are all the products ot blind groping by mutations and of sifting by natural selection. The manner in which the pieces are put together is far from haphazard, however. A nuclear envelope

156

THE AGE OF THE SINGLE CELL

assembles in a strictly reproducible succession of steps that are programmed by the properties of the assembling pieces and by those of the catalysts that mediate the process. The example of the nuclear envelope can be generalized to every complex cytoskeletal structure. Cut off a flagellum, for example, and the whole edifice will grow back from its root by a defined succession of steps. In all cases, structures are the products of spontaneous self-assembly operating according to a program geneti¬ cally inscribed into the properties of the assembling building blocks. How could such molecules as actin, tubulin, clathrin, the nuclear building blocks, and the numerous other cytoskeletal proteins ever have emerged? The key word, I believe, is complementarity, which offers a clue as to what may have been the first critical mutations. Proteins were altered in a way that fitted them with com¬ plementary means of mutual attachment. The first props that allowed cells to grow bigger without collapsing arose in this manner. After that, a long succession of fur¬ ther mutations, each providing some evolutionary advantage, honed these proteins to their present degree of perfection and added hundreds of new proteins that joined with them to build structures of increasing complexity, often endowed with motil¬ ity. Most intermediates in this evolutionary process have been wiped out by natural selection, but molecular kinships that may help in reconstructing the history of the proteins concerned are beginning to be recognized by comparative sequencing. As with every evolutionary problem, the question arises as to what advantages drove natural selection at each tiny step of the long, drawn-out process whereby cytoskeletal proteins were developed and refined. A likely explanation is that the new proteins all played a role in helping the cells to enlarge their volume and expand their surface membrane into a network of increasingly elaborate intracellu¬ lar compartments. Enhanced heterotrophic autonomy may have provided the major selective factor that drove, in mutually supporting fashion, the coevolutionary development of the eukaryotic cytomembrane network and of the cytoskeletalmotor systems. The fact that many new proteins had to be developed to make a eukaryote out of a prokaryote may well explain the exceedingly long time this transformation required.

WHY TWO NUCLEI? SEX AND THE SINGLE CELL Giardia has two apparently identical nuclei. Same size, same shape, same four chromosomes, same genes. Or so it seems. The evidence is not all in yet, but present indications point in this direction. According to Karen Kabnick and Debra Peattie3 from the Harvard School of Public Health, there is a good possibility that Giardia's two nuclei each contain a complete copy of the same genome. In technical jargon, each nucleus is haploid (from the Greek haplous, single), and the cell is diploid (from

THE PRIMITIVE PHAGOCYTE

157

the Greek diplous, double). These are two passwords to the whole of eukaryotic evolution—worth remembering. It is easy to see how a binucleated cell could arise. A cell “forgot” to divide after nuclear duplication, saddling its progeny with two nuclei that were henceforth duplicated and bequeathed in pairs from generation to generation. Or, alternatively, two cells, each with a single nucleus, fused into a binucleated cell by a variant of the bubble trick involving the merger of their peripheral membranes. This phenom¬ enon, which would have been favored by the absence of a cell wall, occurs com¬ monly in nature and is readily provoked. It earned the Argentinian-born British sci¬ entist Cesar Mils’tein and the German Georges Kohler the 1984 Nobel Prize in medicine when they fused an antibody-making cell with a cancer cell. The resulting hybrid cell, which combined the property of making a given type of antibody with the cancer cell’s ability to divide indefinitely, turned into a self-reproducing factory for making monoclonal antibodies on a large scale. Such factories are now at work all over the world and supply invaluable tools for research and medicine.4 Having a second nucleus puts an additional duplication load on a cell. This state would not have been perpetuated had it not entailed an evolutionary advantage. In fact, the benefit of diploidy is enormous and manifests itself whenever a gene suf¬ fers a mutation. Suppose the mutation is harmful. Whereas a haploid cell looses out, the diploid cell has a spare copy of the gene and survives. In the rare case the mutation is beneficial, diploidy is also advantageous. It allows the cell and its prog¬ eny to enjoy the benefit of the mutation and even to explore its further evolutionary possibilities, while the unmutated gene of the pair goes on doing its job. An initially harmful mutation may even be made beneficial in this way by one or more addi¬ tional mutations of the same gene. A consequence of all this is gene diversification. The same gene undergoes different changes in different cells. Thus, many different varieties of the same gene, or alleles—another password—come to be present in the gene pool—still another password—of the species. Diploidy typifies a new kind of evolutionary strategy characteristic of eukary¬ otes. Although bacteria occasionally indulge in gene duplication—and derive evo¬ lutionary benefits from it—their main strategy, helped by rapid multiplication, relies on large-scale genetic experimentation by individuals to take care of almost any contingency. Quantity is exploited, rather than quality. Eukaryotes, blessed and burdened at the same time with an increasingly complex organization and a corre¬ spondingly slower rate of proliferation, were led to evolve a strategy that allowed a similar kind of genetic experimentation while putting greater value on the individ¬ ual. Diploidy was the solution. Two additional developments turned the new strategy into a novel form of the genetic combinatorial game of immense importance. Occasionally, binucleated cells “remembered” belatedly to divide, and the resulting mononucleated cells later rejoined with different mononucleated partners to give rise to binucleated cells hav¬ ing two nuclei of different origin. Because of genetic diversification of individual nuclei, this reshuffling of nuclei often resulted in new combinations of genes that

158

THE AGE OF THE SINGLE CELL

were put to the test of natural selection. The gene pool was stirred and the range of genetic experimentation was expanded. We don’t know whether this back-and-forth movement between diploidy and haploidy ever happens in Giardia, but it is tempt¬ ing to believe that it happened at some stage in the evolution of eukaryotes, as it provides the simplest explanation for the origin of sex. In all forms of sexual reproduction, diploid cells give rise to haploid cells in a special kind of cell division called meiosis. Fusion of two haploid cells then gener¬ ates a diploid cell with its own characteristic set of genes different from that of either of the two diploid parental cells. In the human species, for example, all the cells of the body are diploid, with the exception of the germ cells. Maturation of both sperm cells and egg cells proceeds through meiosis and produces haploid cells. At fertilization, a haploid sperm cell fuses with a haploid egg cell, giving rise to a diploid fertilized egg cell with a unique combination of genes. In its crudest and presumably earliest manifestation, sex amounted to whole nuclei being exchanged between diploid cells. An important refinement was intro¬ duced when the two haploid nuclei of binucleated cells fused into a single diploid nucleus containing all chromosomes in pairs. This merging of the two nuclei into a single one required a major innovation of the mechanism whereby diploid cells gave rise to two haploid cells. Simple division into two mononucleated cells was no longer possible. Two mitotic divisions preceded by a single duplication of the chro¬ mosomes did the job and produced four haploid mononucleated cells from a single diploid cell. This is the mechanism of meiosis, a highly complex process that must have arisen through a long succession of steps, each driven by some selective advantage. We don’t know the details of this evolution but we can guess its main advantage: genetic diversification and the consequent ability to adapt to a variety of circumstances. In a first stage, meiosis allowed individual chromosomes to be exchanged instead of nuclei. This led to a considerable increase in possible combinations. For example, a diploid cell containing four pairs of nonidentical chromosomes can give rise to sixteen different haploid combinations of chromosomes; the number of pos¬ sible combinations is about eight million with a haploid number of twenty-three, as in the human species. Then, the combinatorial range was increased almost to infin¬ ity thanks to crossing-over. In this phenomenon, homologous chromosomes (that is, bearing the same genes, often in the form of different alleles) are closely juxtaposed in a way that permits homologous stretches of DNA to “cross over” reciprocally from one chromosome to the other. The chromosomes that are exchanged after rearrangement by crossing-over are no longer the original parental chromosomes but mosaic chromosomes combining more or less randomly selected pieces of both. This virtually guarantees that each haploid cell formed by meiosis from a given type of diploid cell, and, hence, each diploid cell arising from the fusion of two such haploid cells, has a unique genetic make-up, except when genetic diversifica¬ tion is hampered by inbreeding.

THE PRIMITIVE PHAGOCYTE

159

The eukaryotic form of sex is very much superior to bacterial conjugation. It has given eukaryotes an enormously powerful means of diversification and adaptation, which accounts for much of Their variety and success. It is interesting that sexual reproduction is resorted to by primitive protists only in times of crisis. This is in keeping with a general feature of evolution, which long ago invented the “if it ain’t broke, don’t fix it” principle—not by applying common sense, of course, but because of the simple fact that mutations are rarely beneficial when all goes well. As long as organisms are adapted to their environment, evolution is largely conser¬ vative. Cells multiply by simple division, perpetuating the same genome. But let the survival of the cells be endangered by some environmental upheaval and they suddenly go into a frenzy of sexual debauchery, which, interpreted in anthropomor¬ phic evolutionary terms, amounts to a frantic search for a genetic combination bet¬ ter adapted to the new conditions. For single cells, sex is an emergency measure, not a piece of cake.

Chapter 17

The Guests That Stayed

With

the

emergence of the primitive phagocyte, the major part of the

prokaryote-eukaryote transition was accomplished. Considering the many innova¬ tions needed for this development, the uptake and adoption of endosymbionts may be seen as an almost banal event. Yet it was of paramount importance for future evolution. With rare exceptions, today’s eukaryotes all belong to the post-endosymbiont era. There may be a good reason for this, possibly connected with oxygen.

AN AGE-OLD BATTLE The possession of flagella by Giardia tells us that the primitive phagocyte was a motile, fully emancipated cell that had long left the shelter of the bacterial colonies within which it supposedly was born. Perhaps it tended to browse around the rich and easily accessible food supply offered by its erstwhile abode, but it could also have taken advantage of its freedom to move out into any stream, lake, sea, or ocean where bacteria were present. Quite possibly, it spread in different directions and diversified into a variety of species adapted to different environments. Products of this early diversification that have survived to this day include the diplomonads, the microsporidia, and, perhaps, other members, still awaiting discovery, of the large, incompletely inventoried group of protists. It is likely that our distant eukaryotic ancestor had improved its chances of heterotrophic survival by acquiring some of the properties that help phagocytes today. It probably possessed chemotactic surface receptors sensitive to certain types of molecules and connected to the flagellar apparatus in a manner that made the cell move toward a potential food supply and away from noxious substances. Most likely, it also had endocytic receptors to help it catch and engulf its prey. Perhaps, like our own white blood cells, it complemented the digestive enzymes of its lyso-

THE GUESTS THAT STAYED

161

somes with special killing agents. It may even, like some protists today, have had stinging tentacles with which to stun its victims by means of exocytized toxic chemicals.

0

Predictably, bacteria did not simply submit to their phagocyte-inflicted fate. Thanks to their remarkable potential for meeting all sorts of contingencies— remember the antibiotics—they no doubt evolved a number of countermeasures similar to those practiced by pathogenic bacteria today. Some, like the famous agents of bacterial pneumonia, may have evaded detection and capture by modify¬ ing their wall. Others, like the “streps” (streptococci) and “staphs” (staphylococci) that visit upon us a variety of unpleasant infections, may have responded to attack by the release of toxins that injure membranes, thereby escaping after capture and killing their catcher at the same time. Yet others, like the related agents of tubercu¬ losis and leprosy (mycobacteria), may have evolved a strategy of survival within endosomes or lysosomes, multiplying inside these membrane-lined pockets to the point of causing their host cells to enlarge enormously and eventually disintegrate. Others may have combined the escape and survival strategies by first breaking open the surrounding membrane of their endosomal or lysosomal prison and then proliferating in the cytosol. Occasionally, a stalemate was reached in this perpetual warfare, a sort of truce or mutual nonaggression pact, in which the captured bacteria and their captor spared each other. Such situations, when beneficial to both captor and prisoner, were favored by natural selection and evolved into lasting relationships. The captor became host and the victim guest. Many cases of endosymbiosis are known today and we may take it that many were established in those days when the first eukary¬ otic phagocytes started to roam the world in pursuit of bacterial prey. There is, however, a strange discrepancy. If our timing is correct—huge uncertainties affect all such reconstructions— primitive phagocytes existed before two billion years ago, when Giardia s distant ancestor branched from the main eukaryotic line, whereas lasting endosymbionts were adopted only about 1.5 billion years ago.1 There is thus an immense gap of several hundred million years between the time cells first became capable of cap¬ turing endosymbionts and the time when lasting endosymbionts were actually adopted. Rather than dwelling on this discrepancy in unprofitable speculation, I wish simply to point to a coincidence that may or may not be significant. The first permanent endosymbionts to be adopted were oxygen-utilizing bacteria, and their adoption coincided roughly with the great oxygen crisis. Add to this the fact that the primitive phagocyte most likely was anaerobic and, perhaps, poorly equipped to deal with oxygen, and the possibility comes up that most of the primitive phago¬ cyte’s descendants, with rare exceptions, such as Giardia"s ancestor, fell victim to the oxygen holocaust, leaving as main survivors those that were rescued by oxy¬ gen-adapted endosymbionts. We have no proof for this, but it is an attractive hypothesis. Descendants of these life-saving guests include the mitochondria and, perhaps, the peroxisomes.

THE AGE OF THE SINGLE CELL

162

MITOCHONDRIA: THE CELL’S POWER PLANTS Mitochondria (singular, mitochondrion, from the Greek mitos, thread, and khondros, grain) are conspicuous particulate components of the great majority of protists and of all plant, mold, and animal cells. Their shapes vary from spherical to fila¬ mentous. Their sizes, on the order of one twenty-thousandth of an inch, recall that of their bacterial ancestors. Their numbers may reach several thousand per cell. They are surrounded by two membranes, of which the inner one is pleated by ridge¬ like infoldings, or cristae. This inner membrane, which is derived from the cell membrane of the bacterial ancestor, is crammed with oxygen-linked respiratory chains that generate ATP by way of a protonmotive force. The inside, or matrix, of the organelle contains powerful metabolic systems that break down a great variety of substances to provide the electrons that are fed into the respiratory chains. The outer mitochondrial membrane, which is relatively porous, probably originates from the outer membrane of the (gram-negative) bacterial endosymbiont or, less likely, from the membrane of the endocytic vesicle within which the ancestral bac¬ terium was initially caught. Mitochondria are the main sites of oxygen utilization and metabolic ATP production in all aerobic eukaryotic cells. They are the power plants of the cells. According to sequencing results, mitochondria share a closest common ancestor with a group of present-day aerobic microbes known as purple nonsulfur bacteria. When first caught, these ancestral organisms became established in the cytoplasm of their captors, which supplied them with plenty of food in return for being kept largely oxygen-free. For this relationship to last, the proliferation of the bacterial settlers had to be adjusted to the slower rate of reproduction of their hosts. How this happened in the short run is not known. In the long run, the problem was solved by the transfer of endosymbiont genes to the host-cell nucleus. This phenomenon occurred on a remarkably massive scale, to the point that only a few of the original bacterial genes are left in mitochondria today. These remnants of a past autonomy have fortunately survived, along with appropriate replication, transcription, and translation machineries, to provide us with unmistakable proof of the bacterial ori¬ gin of mitochondria. In keeping with this ancestry, the mitochondrial genome is contained in a circular, relatively unstructured chromosome of characteristic bacter¬ ial type, and the mitochondrial ribosomes also have several bacterial properties that clearly distinguish them from the ribosomes present in the surrounding cytoplasm of the same cells. There is nothing very remarkable about the actual transfer of DNA to the nucleus. Such a phenomenon takes place routinely in transfection, a genetic manip¬ ulation in which DNA is introduced into the cytoplasm of cells by means of a microneedle or otherwise. This DNA readily becomes integrated within the nucleus in a manner such that the foreign DNA is replicated synchronously with the nuclear

THE GUESTS THAT STAYED

163

DNA, as well as transcribed into messenger RNAs that are correctly translated in the cytoplasm. We may take it that the same would have happened to DNA released into the cytoplasm from endosymbionts, following an injury to the endosymbionts, for example. However, the proteins encoded by the transferred genes would then have been synthesized in the host-cell cytoplasm, where they would have been of little use. To perform their functions, these proteins needed to be translocated into the endosymbionts, a phenomenon that required some important innovations. Today, mitochondrial proteins made by cytoplasmic ribosomes are translocated posttranslationally into the organelles by complex, energy-dependent machineries present in the mitochondrial membranes and capable of recognizing special target¬ ing sequences in the proteins. Similar machineries exist in the bacterial cell mem¬ brane and (originally brought in by infoldings of this membrane) in certain parts of the eukaryotic cytomembrane network. The mitochondrial system presumably developed from one of these machineries, but with an adaptation to a different kind of targeting sequence. Two facts render this development a little less improbable than it appears. First, there was no time pressure. While evolution was playing with all kinds of muta¬ tions, enough normal endosymbionts still endowed with the transferred gene remained to save the population from extinction. Next, once the right tandem of translocation machinery and targeting sequence had been established for one pro¬ tein, there was need only for the same targeting sequence to appear in the other pro¬ teins, an event that could have happened by mutations or by transposition of the corresponding DNA sequence and for which, again, plenty of time was available. To paraphrase two aphorisms, there was safety in numbers and the first step was the hardest. However, there remains to be explained why gene transfer from endosymbiont to nucleus occurred on such a large scale and why endosymbionts still in pos¬ session of transferred genes were eradicated in favor of those that had lost the genes. Strong evolutionary advantages obviously drove the genetic subversion of the endosymbionts by their hosts. The most powerful such advantages probably resulted from having the genes gathered together in a central, specially equipped location, where concerted replication could take place, a variety of genetic rearrangements could occur, transcription could be regulated, and RNA products could be processed. An additional benefit could have been that the unencumbered endosymbionts were able to devote themselves more fully to their main tasks of freeing their hosts from oxygen and supplying them with ATR While they were playing these evolutionary games, mitochondria even indulged in the almost unique luxury of tinkering with the genetic code. These mitochondrial deviations occurred late, as they are not the same among plants, animals, and molds and even differ among some species of animals or of molds. They were probably made possible because the number of genes concerned was small enough to allow adaptation to a different genetic language. Mitochondria have elaborate respiratory chains, adapted to retrieve a maximum of energy in usable form from the flow of electrons passing through them. So have

164

THE AGE OF THE SINGLE CELL

their closest bacterial relatives and so, presumably, had the common ancestor of both. This quality fits the present role of mitochondria in eukaryotic cells, most of which have become vitally dependent on mitochondrial respiration and an adequate provision of oxygen for their energy supply. This kind of sophistication, however, has the appearance of a late product in the evolutionary adaptation to oxygen. If aerobic endosymbionts indeed rescued eukaryotes from death by oxygen, one would expect the first rescue operation to be accomplished by a more primitive form of aerobic microorganism. There is an attractive candidate for the role of descendant horn such an organism. It is called microbody by cell morphologists and peroxisome by biochemists, and it is present in the vast majority of cells that contain mitochondria, whether in plants, molds, or animals.

PEROXISOMES: PROTECTORS AGAINST OXYGEN TOXICITY Peioxisomes are particulate entities somewhat smaller than mitochondria and sur¬ rounded by a single membrane unrelated to the general cytomembrane system and possibly inherited from an endosymbiotic ancestor. They contain a variety of meta¬ bolic systems and they detoxify oxygen and some of its harmful derivatives, espe¬ cially hydrogen peroxide, as their name indicates. Unlike mitochondria, they do so in a manner that is strictly unproductive in terms of energy retrieval, resembling some primitive aerobic bacteria in this respect. Their proteins are made in the cyto¬ plasm and transferred posttranslationally to the particles by a system that relies, like that of authentic endosymbiont descendants, on the recognition of specific targeting sequences. The snag is that peroxisomes contain no trace of a genetic system. This in no way invalidates an endosymbiont origin. If mitochondria lost more than 99 percent of their genes to the nucleus, the older peroxisomes could very well have lost 100 percent. However, without the remnants of a genetic system, the case for an endosymbiont origin becomes very much weaker. In addition, sequencing results have so far provided little evidence of a kinship with bacteria. The matter is left open. If peioxisomes preceded mitochondria as protectors against oxygen toxicity, one may wonder why they were not subsequently eradicated after the more efficient mitochondria were adopted. A likely answer to this question is that peroxisomes, in the course of evolution, came to accomplish some vital functions, unrelated to oxy¬ gen detoxification, that neither the host cells nor the mitochondrial precursors could fulfill. Human genetic pathology does, indeed, support this possibility. A number of seveie inborn deficiencies ol lipid metabolism, some of them resulting in the early death of afflicted infants, have been identified as peroxisomal defects. One such disease, adrenoleukodystrophy (ALD), characterized by an inability to break down some special fatty acids, has been widely publicized by the film Lorenzo’s Oil.2

THE GUESTS THAT STAYED

165

CHLOROPL ASTS: THE EUKARYOTIC LINK TO THE SUN 0

After mitochondria were adopted as regular components of virtually all eukaryotic cells, a second major implantation of bacterial endosymbionts took place. More correctly, it was a wave of such implantations, since there is evidence that the same phenomenon occurred several times. In all cases, the guests were cyanobacteria, that is, representatives of the more advanced, oxygen-producing, phototrophic bac¬ teria. The hosts were various eukaryotic cells, all well provided with peroxisomes, mitochondria, and perhaps other systems to help them resist the production of the toxic oxygen in their very cytoplasm. Without such equipment, phagocytes would have been unable to host cyanobacteria. The engulfed cyanobacteria evolved into the characteristic organelles of phototrophic eukaryotes, the chloroplasts. The protists that adopted them turned into various types of green, red, or brown unicellular algae, of which one group was to give rise later to all green plants. Lines that did not adopt chloroplasts led, besides a number of protists, to all molds and animals, which were allowed to maintain the heterotrophic way of life of their ancestors thanks to the luxuriant proliferation of their distant phototrophic relatives. Distinctly larger than mitochondria, chloroplasts are surrounded by two mem¬ branes and filled with membranous stacks that contain the phototrophic machiner¬ ies. These stacks are related to similar formations present in cyanobacteria. The matrix of the organelles, which is derived from the cytosol of the cyanobacterial ancestor, contains a number of metabolic systems, most prominently the key enzymes involved in the assimilation of carbon dioxide, the hallmark of autotrophy. Chloroplasts have the main characteristics of endosymbiont descendants. Like mitochondria, they have a rudimentary but active genetic system, only with more original genes left, in conformity with their younger age. They obey the universal genetic code. Most of their proteins are made in the cytoplasm and taken up posttranslationally through the mediation of specific targeting sequences. Their kinship with cyanobacteria is supported by sequence homologies.

OTHER POSSIBLE ENDOSYMBIONTS The possibility has been evoked that other components of eukaryotic cells besides mitochondria, peroxisomes, and chloroplasts may be derived from endosymbiotic bacteria. Such an origin has been postulated for hydrogenosomes, which are mem¬ brane-bounded, cytoplasmic organelles about the size of mitochondria, uniquely characterized by the ability to produce molecular hydrogen.3 Discovered by Miklos Muller, from the Rockefeller University in New York, in a special group of anaero¬ bic protists called trichomonads, which are parasites of the genital tract in humans

166

THE AGE OF THE SINGLE CELL

and some animals, hydrogenosomes have been detected also in a number of other protists unrelated to trichomonads and in some fungi. Hydrogenosomes have the main properties ot endosymbiont-derived organelles, except that, like peroxisomes, they lack evidence of a genetic machinery. The biological distribution of these intriguing organelles, although highly restricted, suggests that they may have origi¬ nated more than once. Lynn Margulis4 has proposed that flagella and, therefore, the whole microtubu¬ lar cytoskeletal system were brought into eukaryotic cells by flagellated bacteria belonging to the group of spirochetes (which includes the causal agent of syphilis). Some evidence, including the possible association of DNA with centrioles (eukary¬ otic components derived from flagellar roots), has been adduced in support of this hypothesis. Interpretation of the data remains, however, uncertain. It has been men¬ tioned before that bacterial and eukaryotic flagella are totally unrelated chemically. No evidence to the contrary has yet been obtained. The suggestion has also been made that the eukaryotic nucleus, and with it the whole DNA-based genetic system, may have been imported with an engulfed bac¬ terium.'’ This hypothesis implies the existence of a primitive phagocyte using an RNA-based genetic system. It seems to me difficult to visualize how the resulting cell could have simultaneously handled an RNA and a DNA genome.

A LAST LOOK BACK Among the profusion of prokaryotic branches that have sprung from the common ancestor to fill every niche available on our planet, the line leading to eukaryotes stands out as a towering trunk, rising high in solitary splendor before suddenly breaking out into a canopy of luxuriant ramifications that dwarf and overshadow the variegated spread below. One is left with the impression of something unique, almost uncanny, an aberrant growth that developed among millions of “normal” offshoots as a result of some extraordinary combination of circumstances or, per¬ haps, a unique chance event. This impression could be misleading. The lonely eukaryotic tree started as a small bush like all its prokaryotic rela¬ tives. Its growth was far from unerringly straight. Most likely, its real shape is gnarled and twisted, knotted by the stubs of numerous abortive growths and shriv¬ eled limbs. Like all other evolutionary processes, the transition from prokaryotes to eukaryotes was groping and exploratory, with each advance selected from many attempts that have left no trace. However, as choices were made or, rather, imposed by environmental selective forces, the range open to further advances became increasingly narrow. Evolution had to tinker—to cite an expression coined by Fiangois Jacob6

with what it had available and it had to await a favorable muta¬

tion within the limits of previous commitments. There was no question of starting a brand-new tack.

THE GUESTS THAT STAYED

167

Unfortunately, no record is left of the long pathway that converted a prokaryote into a large, mobile, nucleated, phagocytic cell. There is hope, however, that some evidence may be uncovered in the future. Strange life forms may still await isola¬ tion and characterization in the rich world of unicellular organisms. Of all the many changes that marked the appearance of the ancestral phagocyte, those that led to the development of a cytoskeleton and related motor systems prob¬ ably played the most decisive role. These parts were needed to support the forma¬ tion of an intracellular cytomembrane system and demanded a large number of genetic innovations. We don’t know how the new structural proteins emerged, but we may be sure that they did not do so by the stroke of a magic wand. Their birth was slow, progressive, and stepwise. A major directing factor in their evolutionary shaping was their ability to join together into higher-order structures by self-assem¬ bly, a process based itself on chemical complementarity. We have already encoun¬ tered chemical complementarity as the property underlying base pairing and many other phenomena. We now find it to be the key to the self-assembly of cell struc¬ tures, mostly from protein building blocks. Because of the rich array of chemical groupings offered by the twenty amino acids that compose proteins, opportunities for protein-protein associations were almost limitless and depended only on some chance mutation in order to materialize. It is noteworthy that the two most important eukaryotic structural proteins, actin and tubulin, both display complementary regions on the same molecule, so that self-assembly can take place reversibly from a single kind of building block.7 The polar complementary regions allow end-to-end associations of indefinite length, whereas the lateral complementary regions determine the three-dimensional orga¬ nization of the resulting threads as either double-stranded filaments or hollow tubules made of thirteen threads. No evidence of sequence analogy between the two molecules has been uncovered so far. It would seem, therefore, that strong selective advantages favored the development of protein molecules with the ability to bind to their own kind. Subsequently, other molecules appeared that could attach to the first structures to shape more complex assemblages or to provide them with motility. No reliable clue as to the origin of these two key proteins has yet been found in the prokaryotic world. Nor are bacterial proteins known that display a similar pair of complementary regions. Perhaps these are simply cases of lack of detection due to incomplete sampling. Another possibility is that the kind of mutations that give rise to such arrangements are very rare events, which happened to take place only in the eukaryotic line. However, the fact that they happened twice in this line argues against this possibility. A more likely explanation is that bacteria have no use for self-assembling proteins or may even be hampered by them, so that the relevant mutations, when they happened, were rejected by natural selection. Only in the spe¬ cial case of a naked, sheltered heterotroph, provided with full opportunities for cel¬ lular enlargement and membrane expansion, did the mutations find fertile ground for positive selection and further improvement along the long road that led to actin

168

THE AGE OF THE SINGLE CELL

and tubulin, those masterpieces of protein engineering. This road must indeed have been long, considering the degree of perfection attained by the two proteins. It is also a typical example of progressive evolutionary narrowing. Actin and tubulin are both highly conserved proteins displaying closely similar structures throughout the eukaryotic world. This means that they were completed, with virtually no room for further improvement, two billion years ago or earlier. Actin and tubulin are only the two most remarkable products of this extraordi¬ nary evolutionary adventure, which saw the birth of many other proteins, all unknown in the prokaryotic world, that provided the structural props and motile machineries of the first eukaryotes. Probably the main selective forces were the same in all cases and related to the improvement of the phagocytic way of life. The final prize, emancipation, was probably long in coming and achieved only thanks to the remarkable constancy, possibly extending over several hundred million years, of the physical and chemical conditions provided by the environment within which this epoch-making metamorphosis took place. For all we know, there may have been many other trials in the same direction that eventually aborted because envi¬ ronmental constraints did not allow them to come to fruition. Some may even have been successful but produced lines that became extinct for one reason or the other. In the second phase of eukaryotic evolution, after the primitive phagocyte had emerged, this organism was able to diversify and invade a variety of environments, though apparently without undergoing major evolutionary changes, until the wave—possibly precipitated by the great oxygen crisis—of endosymbiotic adop¬ tions that gave rise to modern eukaryotes. This is a typical feature of evolution. A given group may remain static for a prolonged period of time, marked only by the kind of point mutations that do not influence the performance of the affected mole¬ cules but provide valuable signposts of evolutionary distance. Then, rather sud¬ denly, most often as a result of climatic or other environmental changes, a fairly rapid transformation occurs, giving the impression of an evolutionary jump, though only in relation to the long static period that came before. The pace of evolution is variable but not discontinuous. After the appearance of the first endosymbiont-containing protists, evolution once again settled into a relatively static mode, engaging mostly in diversifica¬ tion

endless variations on the same basic themes of oxygen-producing photo¬

trophy and aerobic heterotrophy, without a truly novel theme emerging. Then some eukaryotic cells “discovered” the advantages of getting together and pooling efforts. Why it took them so long to make this discovery is not clear. An enhanced interest in sex could be, at least, part of the answer, together with some major envilonmental change that made intercellular cooperation advantageous. This question will be examined in the next part.

PART V

THE AGE OF MULTICELLULAR ORGANISMS

Chapter 18

The Benefits of Cellular Collectivism

Cells

remained

single for about three billion years. Bacteria still are

today; they do sometimes form colonies—remember the stromatolites—but not true organisms.1 This may be related to their “selfish” mode of life, geared entirely to producing as many progeny as possible in as short a time as possible. Even eukaryotic cells have clung to singleness for hundreds of millions of years. Cells endowed with all eukaryote attributes, including endosymbionts, have been around for well over one billion years. Yet there is no trace of multicellular life before 600 to 700 million years ago. Unicellular protists are still abundant in the present-day world. What prompted some eukaryotic cells to join is not known, except in a general sense. We may take it that cells first got together as a result of chance mutations that favored their association, and that they stayed together because they were reproductively more successful as a group than single. Once they took hold, the advantages of collectivism were swiftly exploited further by evolution, to generate the rapidly expanding and diversifying worlds of plants and animals. Why was this discovery not made before? And why was it made when it was, and then almost simultaneously by autotrophs and heterotrophs? It is possible that some major envi¬ ronmental change made cooperative behavior more advantageous, perhaps by putting a premium on sexual reproduction. This possibility is supported by the behavior of slime molds, organisms that may be viewed as intermediate between unicellular and multicellular.

SLIME MOLDS: AN INSTRUCTIVE EXAMPLE The most ancient attempt at eukaryotic, heterotrophic cooperation on record was performed by the remote ancestors of organisms named slime molds, or myx-

172

THE AGE OF MULTICELLULAR ORGANISMS

omycetes. This is a misnomer: These organisms have nothing to do with molds or mycetes. Neither are they plants or animals in the usual sense of these words. T hey are the survivors of an evolutionary experiment made more than one billion years ago that never really caught on. They do, however, convey an interesting message. Slime molds consist of unicellular, heterotrophic protists similar to amoebae. Like amoebae, these organisms wander about in search of prey, which they catch by phagocytosis and digest intracellularly. Let the food supply become scarce, how¬ ever, and the cells exchange a chemical signal—the agent is cyclic adenosine monophosphate (cAMP), a universal chemical transmitter derived from ATP—that causes them to aggregate into a single mass. This collective then starts crawling, leaving behind a slimy trail, and progressively builds itself into an erect structure called a fruiting body. This structure produces, sometimes through a sexual process, a special kind of protected cells, called spores, that are shed and lie dormant as long as conditions remain unfavorable. When circumstances improve, the spores mature into amoeba-like forms, which resume their unicellular mode of life. The formation of spores is a common phenomenon in the unicellular world. Many bacteria and protists react to adverse environmental changes by encasing themselves within a protective shell and entering a state of metabolic torpor, await¬ ing “better days.” Slime molds are the first example of cooperative sporulation, a phenomenon exhibited by many plants and fungi. Slime molds also illustrate a mechanism of interest with respect to the emer¬ gence of animals, though in a different context. Upon exposure to cyclic AMP, the unicellular forms of the organism express new surface molecules with mutually complementary lock-and-key arrangements that keep cells stuck to each other after a chance encounter. The cells are also held together indirectly by means of surface receptors that bind them to the viscous, extracellular material they secrete. This “slime” serves as a glue, as a carpetlike substrate, and as a recognition trail. Animal cells are likewise held together by surface adhesion molecules that join the cells to each other (cell adhesion molecules, or CAMs) and to extracellular scaffoldings (substrate adhesion molecules, or SAMs). A third lesson we learn from slime molds is the role of sexual reproduction as an emergency measure. In the history of multicellular eukaryotes, this mode of repro¬ duction progressively became a major factor of evolutionary resilience and diversi¬ fication.

THE IMPORTANCE OF SEXUAL REPRODUCTION Bacteria engage in conjugation and genetic recombination. Real sex, however, with its systematic alternation between diploidy and haploidy, is a typical eukaryotic prerogative, which was probably first engaged in by the primitive phagocyte. The

THE BENEFITS OF CELLULAR COLLECTIVISM

173

main advantages of sexual reproduction were considered in chapter 16. What was not considered in that chapter is the influence of sexual reproduction on evolution¬ ary mechanisms. When cells multiply by simple division, entire genomes are reproduced with, occasionally, the appearance of a mutant combination that becomes, itself, subject to reproduction. Selection operates among these different forms of the same genome. Some may continue to diverge side by side. Things are more complex in sexual reproduction. Mutant genes become associ¬ ated with different gene combinations at each generation. Their evolutionary effects must be assessed x>n a statistical basis, by their ability to spread into the gene pool of the population. For this reason, two evolutionary lines can separate only in reproductive isolation, that is, if they are unable to interbreed. A special discipline, population genetics, has been developed to deal with these problems. Its method¬ ologies are too complex to be elaborated on here, but its existence deserves to be mentioned, as I shall make little reference to it in the simplified accounts given in the coming chapters.

PRINCIPLES OF CELLULAR COLLECTIVISM A central tenet of modern Darwinian theory asserts that evolution proceeds by ran¬ dom mutations, the effects of which are screened by natural selection. All the find¬ ings of molecular biology support this view. This does not mean that evolution is haphazard. Running through the processes of multicellular complexification are a few unifying threads: association, differentiation, patterning, and reproduction.

Association Every cell is born by division, next to a sister cell. If something tends to keep the two cells together, they will stay together. This will happen if sister cells stick to each other or if they remain within a shared housing. When each of the two sister cells divides, the same phenomenon keeps the resulting foursome together. Repeti¬ tion of this process gives rise to a colony of increasing size. Since plenty of mutations may occur that favor or impede the associative behav¬ ior of cells, it is left for natural selection to weigh the advantages of congregation against its disadvantages. The main drawback: Cells are likely to have less access to sources of nutrients and energy when grouped together than when isolated. On the plus side are better protection against predators and environmental injuries and, especially, the benefits of cooperativity. Colony growth cannot be indefinite, how¬ ever, and must at some stage give way to colony reproduction.

174

THE AGE OF MULTICELLULAR ORGANISMS

Differentiation True colonies composed of the same kind of cells are a rarity. Association gains a particularly powerful advantage from differentiation, whereby genetically identi¬ cal cells become different by no longer expressing the same genes to the same extent. The seeds of differentiation lie in gene regulation, which governs many adaptive behaviors. The way in which bacteria adapt to milk sugar by turning on the genes coding for the necessary enzymes is a typical example (see chapter 14). Regulation by transcriptional control of gene expression also takes place in multi¬ cellular eukaryotes. In puberty, for example, what causes a girl to develop breasts or a boy to grow facial hair is the transcription of certain genes in the cells con¬ cerned, induced by the hormones whose secretion sets off puberty. Transcriptional control of genes is particularly important in development. It explains why cells possessing the same set of genes may be very different—liver cells, muscle cells, nerve cells, and so on, or, in plants, root cells, bark cells, leaf cells. It all depends on which genes they express. These effects are mediated by special proteins, called transcription factors, endowed with the ability to interact with certain specific regions of DNA. The genes coding for such transcription fac¬ tors are designated regulatory genes, as opposed to the genes that code for enzymes or structural proteins, which belong to the group of “housekeeping” genes. Differentiation allows cellular specialization and, thereby, division of labor among members of a collective; it is the secret of cellular cooperativity and evolu¬ tionary complexification. Differentiation runs through all the ramifications of the tree of life. Between seaweed and magnolia, between sponge and eagle, one impor¬ tant difference is the number of distinct cell types that compose the organism. But this is only part of biological diversity. Another is patterning.

Patterning The body of a human adult contains several trillion cells but only about two hun¬ dred cell types. Essentially the same cell types serve to build the bodies of a mouse or a whale, or even, with few differences, the bodies of a frog or a fish, just as the same types of bricks and planks may serve to build different dwellings, from cot¬ tages to mansions. The paramount significance of patterning is clear. If we wish to understand evolution, we must pay special attention to what the American biologist Gerald Edelman has called topobiology,2 the study of the mechanisms that cause differentiated cells to assemble into characteristic three-dimensional patterns. Since evolution proceeds by way of genetic alterations, the changes that produced a mouse, a whale, or a human from their common mammalian ancestor, or even a fish, a frog, or a mammal from a primitive vertebrate, must largely result from mutations affecting pattern-controlling genes.

THE BENEFITS OF CELLULAR COLLECTIVISM

175

Reproduction All multicellular organisms arise from a single cell—spore or fertilized egg— genetically programmed to enact with great precision a scenario of coordinated divisions, differentiations, and patternings resulting in organisms similar to the par¬ ent organisms and capable of perpetuating the species in a similar way. The same scenario is re-enacted at each generation. This reproductive behavior holds several implications basic to the process of evolution. First, the target of mutation is the progenitor cell. Only mutations affecting pro¬ genitor cells are ^relevant to the evolutionary fate of multicellular organisms. A somatic mutation (from the Greek soma, body) may have a major effect on the via¬ bility of the affected organism but, not being transmissible, cannot affect the organ¬ ism’s progeny. Second, the target of selection is the organism. A progenitor cell that has under¬ gone a mutation must produce a complete organism if the mutation’s effect on via¬ bility and reproductive success is to be evaluated by natural selection, at least if selection is to be positive. Negative selection can occur any time after fertilization. Third, the developmental blueprint of an organism is written into the genome of progenitor cells. To have an evolutionary impact, progenitor cell mutations must affect genes that control development, that is, regulatory genes. And finally, evolution operates within the constraints of an existing develop¬ mental blueprint. The more complex this blueprint, the more severe the constraints. With only a few lines sketched out, a drawing still has the potential to become a landscape, a still life, or a nude, depending on the whim of the artist. As more details are put in, the commitment becomes increasingly stringent. This rule, allimportant to our understanding of multicellular evolution, explains why only a small number of distinct body plans, all dating back to early stages of evolution, underlie the profusion of different organisms that have arisen.

Chapter 19

The Greening of the Earth

One billion years ago, the continents were barren expanses of rock and lava, deserts baking in the sun by day and freezing by night, rarely refreshed by rainfall and unable to retain moisture for lack of topsoil.1 In contrast, the oceans were filled with all kinds of unicellular life. Bacteria were abundant. So were uni¬ cellular eukaryotes, which already had diversified into a variety of phototrophic and heterotrophic species, many of which had developed a sexual mode of repro¬ duction as an alternative, under special conditions, to their usual vegetative way of multiplying by division. In this watery laboratory, protists formed all sorts of asso¬ ciations, most of which failed to survive. A few of these associations turned out to be advantageous and developed further. Multicellular eukaryotic forms of life probably arose initially from small clones of cells that remained associated after their production, by successive divisions, from a single parent cell. The cells were held together either by intercellular con¬ nections or by a shared external wall or shell. Roughly speaking, the former mecha¬ nism led to animals and the latter to plants and fungi. This division reflects key dif¬ ferences in lifestyle. The heterotrophic animals had to maintain freedom of movement in order to catch prey, even if this freedom meant greater fragility. The phototrophic plants needed only to catch sunlight (and dissolved mineral nutrients) and could afford to remain immobile, even derived an advantage from being immo¬ bilized in a favorable location. Fungi, which developed a scavenging form of het¬ erotrophy based on the breakdown of dead organisms by means of secreted diges¬ tive enzymes, were able to forsake mobility for the advantages of a protective coating. Because of such fundamental differences, these three kingdoms followed very different evolutionary pathways. Most easily reconstructed is the early history of plants,2 because species that may be representative of successive evolutionary stages still exist today. There is danger in this extrapolation from present to past. Extant alleged “missing links” all evolved over long periods and may in no way resemble their distant ancestors. One might even say that they could not possibly resemble their ancestors. Otherwise,

THE GREENING OF THE EARTH

177

why were they not wiped out by natural selection? This is a problem, though not an insurmountable one. Evolutionists no longer tend to view change as a necessary concomitant of evolution, but as something that happens only if enforced by cir0

cumstances, most often an environmental change. If a form of life is well adapted to its surroundings, it may persist unchanged as long as its niche remains unaltered. Even a poorly adapted form may survive indefinitely if competition is weak. The prudent recourse to sexual reproduction on the part of many protists illustrates nature’s inherent resistance to change.

ALGAE AND SEAWEEDS Mementoes of early multicellular plant life are found today in the variegated world of algae and seaweeds, from the tiny organisms responsible for the emerald tinge of many a pond to the thick, brown kelps that cover coastal rocks with glistening manes, undulating with the ebb and flow of assaulting waves, or that used, so the legend goes, to ensnare the imprudent navigators who ventured into the Sargasso Sea. At least three distinct evolutionary lines of algae exist, each related to the endosymbiotic adoption of a different kind of phototrophic cyanobacterium. They are, in order of decreasing ancientness, the red, brown, and green algae. With rare exceptions due to secondary regressions to parasitic life, all are phototrophic and produce molecular oxygen. Their chloroplasts all contain green chlorophylls, but with various amounts of accessory pigments of different colors. There is great diversity of size, shape, chemical composition, metabolism, developmental pattern, and reproductive behavior within each of the three groups. A common feature is the construction of external walls made of carbohydrate poly¬ mers, among them cellulose, a glucose polymer that plays a dominant structural role throughout the whole plant world, and a variety of viscous or gummy sub¬ stances, several of which are used industrially. Whenever you savor an ice cream, there is a good chance that the smoothness caressing your palate is due to alginic acid, a carbohydrate polymer extracted from certain kelps. The morphological organization of multicellular algae is usually simple, consist¬ ing most often of branched filaments, sometimes of flat leaflike sheets, uncon¬ nected by a vascular system. Their most prominent specializations include an anchoring structure, called a holdfast, whereby many seaweeds are attached to solid surfaces; bladders or air sacs, which serve as floats; and primitive sex organs. The reproductive behavior of algae varies along a scale of complexity that is often depicted as a recapitulation of the evolutionary history of reproductive function. All algae have a sexual mode of reproduction, involving the fusion of two hap¬ loid gametes (from the Greek gamos, marriage) into a diploid zygote (from the Greek zygos, yoke). Haploid cells have a single set of chromosomes; diploid cells, two. In the simplest and, probably, most primitive form of sexual reproduction, the

178

THE AGE OF MULTICELLULAR ORGANISMS

two gametes look identical. They may be mobile and depend on flagellar motion to find each other. Or they may lack mobility and be brought together passively by appropriate adaptations of the enclosing walls. At the other end of the spectrum, gametes show extensive sexual dimorphism. One is small and flagellated, like male spermatozoa; the other is large, immobile, and stocked with reserve substances, as are female egg cells. Both kinds of gametes are usually produced by the same plants. These are hermaphroditic, joining the attributes of the Greek god Hermes with those of the goddess Aphrodite. In algae, meiosis, the kind of cell division whereby the number of chromosomes is halved and haploid cells are formed from diploid cells, rarely produces gametes directly. The first haploid cells to be formed, called spores, may go through more or less complex phases of multiplication and development before giving rise to the gametes. In the extreme form of this growth pattern, the organism is haploid in all its stages, with the exception of the zygote, which goes through meiosis immedi¬ ately after being formed. The other extreme, in which the organism is entirely diploid except for the gametes, is also known. In many cases, the situation is inter¬ mediate between these two extremes. A spore develops into a haploid organism, which produces gametes, which fuse into a zygote, which develops into a diploid organism, which produces haploid spores, which start a new cycle. The haploid and diploid organisms often have similar shapes. Known as the alternation of genera¬ tions, this pattern is characteristic of many algae and has become, under innumer¬ able variations, a leitmotiv of plant life.

MOSSES INVADE THE LANDS Simple though they are, algae are perfectly adapted to their aqueous milieu and have thrived in it ever since their emergence. What caused some to move out of their balmy abode to confront the rigors of land? Overcrowding, exclusion by more successful species, excessive grazing by animals are possible explanations, though not very convincing ones. The odds to be overcome were of such magnitude that little short of a life-and-death situation could account for the transition. Most likely, certain bodies of water became cut off from the oceans and slowly dried out, leav¬ ing to survive only those forms of life that succeeded in adapting to increasing dry¬ ness. Adaptation was progressive, following the gradient of decreasing wetness along the coastal edges. The least adapted forms remained nearest to the water; the best adapted survived farthest from it. At first, the plants were still intermittently pro¬ vided with essential water and minerals by tides and waves, so the first hurdle was to avoid desiccation between wet episodes. Plants that acquired an impermeable waxy covering, or cuticle, gained a selective advantage. However, this advantage was curtailed by the demands of nutrition. Modifications that allowed the plants to

THE GREENING OF THE EARTH

179

imbibe mineral-laden moisture from the soil were favored by natural selection. So were openings in the cuticle, the precursors of today’s stomata, that made it easier for the phototrophic cells to absorb atmospheric carbon dioxide and get rid of oxy¬ gen. Also useful were any surface projections that anchored the plants to the ground and prevented them from being blown away from vital moisture by the wind. Some of these appendages doubled as imbibing structures, or rhizoids, prefiguring roots. One last development was needed for plants to become fully established on land. Their reproduction had to be ensured without the participation of aquatic progenitor cells. Alternation of generations provided evolution with the appropriate mecha¬ nism. The haploid Spores developed protective coverings and served as vehicles for aerial dissemination. In the soil, the protected spores could remain dormant until enough moisture was present to trigger germination. The haploid plants arising from germinating spores produced motile male gametes and immobile female eggs in neighboring structures that were kept sufficiently humid to allow the male gametes to swim toward the eggs and fertilize them. The resulting diploid zygotes, after going through an abbreviated developmental phase, then gave rise to haploid spores. Thanks to these adaptations, primitive mosses began to cover the shores with furry, green carpets, extending farther inland as their rhizoids penetrated deeper in the soil to catch water and minerals. Apparently, only the green variety of algae succeeded in colonizing dry land. They did so by way of a smoothly progressive adaptation of their ancestral algal blueprint. We can account for the entire evolutionary sequence by small additions or changes to this blueprint, favored at each step by improved fitness to survive and reproduce on land. The continually receding water line exerted considerable selec¬ tive pressure in favor of these modifications, which would have been of little value in a watery milieu. These facts illustrate the power of environmental factors to influence the direction of evolution and the inherent constraints that force evolution to proceed within the framework of an established body plan. Once a successful strategy for survival has been developed, environmental pres¬ sures relax while intrinsic constraints become more stringent. What then follows is mostly secondary radiation, invasion of an increasing number of ecological niches through an increasing diversification of details. This is why mosses still thrive today, divided into some 15,000 distinct species adapted to a wide variety of cli¬ mates, from tropical to arctic, and clinging to a great diversity of supporting struc¬ tures, from waterlogged bark in a rain forest to the barest of rocks.

VASCULARIZATION, A CRITICAL ACQUISITION Early land plants were mostly confined to humid coastal fringes, leaving large expanses of dry land still barren and open to invasion. Conquest of the ancient

180

THE AGE OF MULTICELLULAR ORGANISMS

deserts was achieved inch by inch by mutant plants that progressively acquired a root system capable of penetrating deeper into the soil and absorbing water and mineral nutrients more efficiently. Several other changes accompanied this devel¬ opment. The body of these plants became polarized into two distinct growth zones: the colorless, subterranean root tips and the aerial, green buds, separated by a sys¬ tem of connecting stems. At the same time, the plants became sensitive to the Earth s gravitational field (geotropism) and tended to adopt an upright position. Finally, and most importantly, they developed conducting channels that allowed water and minerals absorbed from the soil to flow up from the roots to the other parts of the plant, and the organic, photosynthetic products made in the green parts to flow down to the roots and other colorless parts. Thanks to this vascularization, the plants were able to grow bigger while expanding their light-catching, photosyn¬ thetic parts into a ramified system of flattened leaves. A major step in evolution was accomplished. Reproductively, the first vascular plants, like their predecessors, went through alternating haploid and diploid generations and used haploid spores as the means of dissemination, but with a marked shift in emphasis from the haploid to the diploid stage. Whereas the dominant form of mosses was the haploid, gamete-producing form, as in many algae, that of the early vascular plants became the diploid, sporeproducing form. Shed spores, after germinating in the soil, gave rise to mature gametes at the end of an inconspicuous, often subterranean developmental phase. The gametes then fused into the diploid zygote from which the main plant grew, often attaining considerable size. With these developments, the stage was set for one of the most fateful events in the saga of life. About 400 million years ago, the green armies marshaled from the oceans began invading land on a massive scale, helped by climatic and geographic changes that occurred at that time and to which they themselves contributed with the water they pumped out of the soil. The atmosphere became more humid, the rains more abundant, and the soil better able to retain moisture. Bacteria of all kinds accompanied the invaders and were soon followed by the first land-based fungi and animals, further enriching the biotope. The plants grew bigger and developed a tough, polymeric substance called lignin that allowed the building of solid trunks. Trees appeared, up to forty feet high and three feet in diameter. Much of the land turned into enormous tropical swamps, harboring rich vegetation that grew much faster than the heterotrophic organisms that fed on them. The dead remains of these plants were left to accumulate and fossilize, creating the huge deposits of carbonrich material now mined as coal. Hence the name Carboniferous given to the geo¬ logic era, between about 360 and 286 million years ago, when those swamps flour¬ ished. Most of the pioneers in the conquest of land are long extinct. Their closest extant relatives, according to the fossil record, include the horsetails (Equisetum), the lycopods, or club mosses, whose flammable, powdery spores were used in my youth to envelop stage monsters in fiery veils, and, especially, the ferns, of which

THE GREENING OF THE EARTH

181

some nine thousand species are known. These plants are mere shadows of their ancestral glory, surviving relics of a bygone era. What caused their downfall? As in most evolutionary upheavals, changes in geographic and climatic conditions were responsible.

THE PERMIAN CRISIS AND THE FORMATION OF SEEDS ✓

After 50 million years of spectacularly successful development, the great Carbonif¬ erous swamps began to dry out and their forests slowly withered away. Not just the land plants but much of marine life was wiped out as well, in what may well have been the most dramatic mass extinction in the history of life on Earth, the great Per¬ mian crisis (the Permian is the geologic period extending between 286 and 250 mil¬ lion years ago).3 Probably the main cause of this catastrophe was the drifting together of all the lands of the Earth into a single continent, Pangaea. Much of the interior of this mass turned into an immense land-locked desert, like the Gobi Desert today. In addition, the climate became much colder, due, perhaps, to cata¬ strophic volcanic eruptions in what is now Siberia, which obscured the skies and blotted out the sun. A good part of Pangaea was situated over the South Pole and was covered by a thick ice sheet. Glaciers lined its coasts with massive frozen cliffs, which crumbled periodically into huge icebergs that were carried away by the currents to cool the seas right up to the tropics. The water level dropped, and much of the sunlight that fell on the Earth’s surface was lost by reflection. The Earth had entered the harshest ice age in its history. Plants reacted against this cataclysmic situation by replacing spores with seeds as a means of dissemination. Or, more likely, some seed-producing species already existed but did not flourish until circumstances turned this property into a vitally important asset, allowing survival where the spore-bearing plants could not hold out. The transition from spore to seed signaled female emancipation. The first step, already accomplished in certain seedless plants, such as lycopods, was a separation of the sexes at the spore level. Instead of a single kind of spore giving rise to a her¬ maphroditic haploid organism that produced both kinds of gametes, two kinds of spores germinated into two different gamete-producing entities. Large macrospores developed into female organisms, small microspores into male organisms. It was then necessary for the male organism’s sperm cells to seek a female organism to find egg cells to fertilize. Subsequent events took place as with hermaphroditic organisms. The fertilized egg cell developed into an early embryo, which eventu¬ ally became implanted in the soil. This separation favored outbreeding over inbreeding, the main disadvantage of hermaphroditism, and allowed experimentation with all kinds of different diploid

182

THE AGE OF MULTICELLULAR ORGANISMS

genomic formulas. But it decreased the chances of fertilization. The next evolution¬ ary step obviated this drawback by shifting the site of fertilization from the soil to the plant itself. Macrospores no longer were shed to germinate in the soil. They completed their maturation on the plant, in special organs called ovules in which the eggs arose within a cocoon of protective and nutritive structures. Male spores continued to be dispersed as wind-borne pollen grains, which, how¬ ever, were now programmed to pursue their maturation only in a compatible ovule. The haphazard character of this mode of dissemination was obviated by the vast number of pollen grains produced. A pollen grain landing on an ovule matured into a sperm cell, which entered the ovule to fertilize an egg cell. The ensuing zygote developed into an early embryo, up to a stage where further development was arrested and the ovule closed around its occupant. This protected, dormant embryo became the seed, which was shed. After seeds were dispersed, their covering protected the embryos against cold and dryness, awaiting favorable circumstances that induced the embryos to resume development and break out of their protective shells. Reserve substances included within the seeds provided the nutrients necessary to sustain the embryos during the time needed for the first rootlets and leaflets to become functional and support autonomous growth. Such an adaptation would be of little use to a plant living in a swamp but would save the species if geographic and climatic conditions became harsher. Seeds are sturdier vehicles of dissemination than spores; they are capable of resisting extreme physical conditions for months, if not vears or even cen¬ turies—the record, held by some lotus seeds found in a peat deposit in Manchuria, exceeds one thousand years4—until a favorable moment, be it fleeting, arrives to permit germination and implantation. After the first wave of spore-bearing plants was decimated by drought and cold, a second wave, equipped with hardy seeds in lieu of spores, invaded the inhos¬ pitable lands of Pangaea. The extant descendants of this second army include the seed ferns, the palm treelike cycads, the ginkgoes, and, especially, the pines, spruces, cypresses, redwoods, and other conifers. Together, these groups form the superfamily of gymnosperms (from the Greek gymnos, naked, and sperma, seed). Actually, their seeds are hardly naked; they are described as such in contradistinc¬ tion to the angiosperms (from the Greek aggeion, envelope), whose seeds are con¬ tained within fruits.

FLOWERS AND FRUITS: THE CROWNING ACHIEVEMENT Angiosperms are the most advanced and abundant forms of plant life on Earth. They started spreading across the continents about 100 million years ago. By that time, Pangaea had moved northward and broken up into continental masses that

THE GREENING OF THE EARTH

183

were drifting apart toward their present-day positions. Nobody knows how the new plants came about, but we may imagine. One day in those remote times, a seed plant suffered a mutation that caused the leaves around the sex organs to lack chlorophyll and turn white or, perhaps, yellow or pink if they retained some acces¬ sory pigments. The uniformly green landscape of the primeval fields and forests thus became dappled for the first time with bright patches, which acted as beacons for insects that happened to be genetically programmed to move toward light. Because of this fortuitous circumstance, the genetic accident suffered by the plant turned into a benefit. In the course of their visitations, the attracted insects collected pollen grains on their bodies from the male organs of the plant and dropped some grains again on female organs. The pollination record of the mutated plant increased, and so did its reproductive success. The insects also profited from the plant’s mutation, which guided them to nutritious nectar; they proliferated in large amounts. Once initiated, the new evolutionary process moved on ineluctably. Fur¬ ther plant mutations created new shapes and colors and a variety of scents, attract¬ ing all sorts of pollinating insects and also other animals, such as birds and bats. Propelled by the most far-reaching instance of mutually advantageous relationships ever established between plants and animals, a revolution was launched that sprin¬ kled the green expanses of the Earth with countless dabs of color. Flowers were born. The key property of flowers is to develop into fruits. This term encompasses not only the oranges, grapes, apples, plums, berries, and other fruits we call by that name; it also includes nuts, ears of corn, pea-filled pods, and the many winged or fluffy seed-containing gondolas that weeds and trees send out floating in the wind on a summer day. A fruit may be defined as one or more seeds enveloped in a hull—although unfertilized flowers can be coaxed to yield pipless oranges and seedless grapes for the fastidious. The hull, derived from the female part of the flower, distinguishes angiosperms from gymnosperms. It consists of a protective covering and nutritive tissues. It owes its origin to a special process called double fertilization, which is unique to flowering plants. While one male sperm fuses with an egg cell to form the zygote out of which the embryo is to grow, a second sperm cell fuses with a diploid cell in the female part of the flower. The triploid cell aris¬ ing from this second fertilization develops into the fruit’s hull. Such is the basic theme on which evolution has composed an extraordinary number of variations, to the delight of our senses. In this new evolutionary phase, plants with male and female organs grouped in a single flower gained a selective advantage, although this never became the rule. Plants with separated sex organs or even plants with two distinct forms, each bear¬ ing a single kind of sex organ, also exist. Many different kinds of insects and other animals participate in the great pollination game, with flowers displaying an aston¬ ishing array of specialized lures and traps selected to ensure that the pollen reaches its target. Summarizing the evolutionary history of plants on Earth, figure 19.1 shows in

184

THE AGE OF MULTICELLULAR ORGANISMS FIGURE 19.1

A Bird’s-Eye View of Plant Evolution Angiosperms

f Flowers, Fruits-

-> Gymnosperms

Seeds-

> i ci ns, c cc.

Vascularization-

^ Mosses

Land Adaptation, Spores

> unicellular Algae

Association-

\ Ancestral Algae This figure depicts the main steps in the rise of plants in the direction of greater complexity. At each folk, a mutant evolutionary line that underwent the genetic change in body plan indi¬ cated on the left diverges upward from the unmutated line—represented by the arrow curv¬ ing to the right—leading to existing phyla.

highly schematic form how key mutations of “fork organisms” led to significant evolutionary advances, leaving the descendants of unmutated organisms to provide us with some information on the properties of the fork organism, the last ancestor they have in common with the more evolved species.

UNDERGROUND INFILTRATORS When plants started invading land, they were soon followed by a horde of scav¬ engers. Related to the unicellular yeasts, these opportunistic fellow conquerors resembled primitive plants in forming branched, tubular structures, or hyphae, encased by tough carbohydrate polymers, and in reproducing by means of spores; but they were colorless and strictly heterotrophic. Unlike other heterotrophic organ¬ isms, they were immobile and unable to catch prey, relying entirely on a primeval form of extracellular digestion for their subsistence. Clinging tightly to their com-

THE GREENING OF THE EARTH

185

panion plants or to the remains of dead plants, sometimes strengthening their hold by means of tiny rootlets, they attacked these substrates with powerful digestive enzymes, which they secreted from their cells, and they absorbed the soluble prod¬ ucts of this digestion by permeation through their surface. This primitive, apparently precarious way of life has turned into a remarkably successful formula, the prerogative of more than 200,000 species forming the vast group of mycetes (from the Greek mykes, mold), which includes yeasts, molds, rusts, smuts, mushrooms, toadstools, puffballs, and sundry other fungi. Long classi¬ fied as members of the plant kingdom and viewed as originating from degenerate plants that had lost their chloroplasts, mycetes are now considered a separate kingdom, distinct from the plant and animal kingdoms. Contrary to former belief, the mycetes have been found by molecular sequencing to be more closely related to animals than to plants.5 Mycetes are prime scavengers and play an important role in the recycling of bioelements. A number are parasitic, causing a variety of diseases in plants and, less frequently, in animals. Others have long been used for their ability to catalyze various fermentations useful in the preparation of breads, cheeses, and alcoholic beverages. The misadventures, sometimes deadly, of unwary mushroom eaters and the opposite experiences of millions of patients saved by penicillin and other fungal antibiotics are witness to the chemical versatility of mycetes. Many fungi live essentially underground and make their presence known only when they suddenly sprout some reproductive structure that emerges from the soil to disperse its spores. Some have established lasting symbiotic associations with green algae, forming the lichens, one of the hardiest forms of life. The body plan of mycetes has remained simple, made mostly of a network of interconnected hyphae called a mycelium. This network may extend over huge areas of up to several square miles. The nuclei of fungal mycelia are always hap¬ loid, but the cells themselves are often binucleate, as are those of Giardia, the most ancient of presently known eukaryotes. This is due to a delay, often of considerable duration and including a number of cell divisions and other developmental events, separating nuclear fusion from cell fusion in sexual reproduction. In plants and ani¬ mals, the two phenomena follow each other rapidly in the course of fertilization. Nuclear fusion does eventually take place in mycetes, and the resulting diploid nucleus then almost immediately goes into meiosis and forms haploid nuclei. These emerge with a reshuffled genetic content thanks to meiotic chromosome rearrange¬ ments, the main evolutionary benefit of sexual reproduction. The uninucleate, hap¬ loid cells produced by meiosis give rise to spores, which are shed to disseminate the species. Sporulation is the major event in the life of most mycetes. It is accompanied by the development of special structures, of which mushrooms are the most spectacu¬ lar examples. In some cases, these organs are built to forcibly eject and spread the spores around. The most famous of fungal spores was released by a common mold, Penicillium notatum, on a September morning in 1928, and landed on a bacterial

186

THE AGE OF MULTICELLULAR ORGANISMS

culture in the laboratory of a Scottish microbiologist, Alexander Fleming, at St. Mary s Hospital in London. The spore developed into a fluffy, greenish colony that killed all the microbes around it, creating a clear circular zone that, fortunately, was noticed by Fleming. The outcome, fifteen years later, was the miracle drug peni¬ cillin, thanks to the persevering efforts of an Australian pathologist, Howard Florey, and an emigre German chemist, later naturalized British, Ernst Chain; thanks also to the special circumstances created by World War II, which justified an extraordi¬ nary outlay of energy and money that might perhaps never have been expended under peaceful conditions.6

Chapter 20

The First Animals

At the time unicellular phototrophic algae started to assemble into the first primitive seaweeds, heterotrophic protists were also led by the vagaries of mutation to experience the virtues and drawbacks of multicellular association, leaving it to natural selection to pronounce the final verdict. Because of the overwhelming need for food that dominates heterotrophic life, the selective advantages that drove ani¬ mal evolution were different from those that propelled the evolution of plants, depending mainly on improved feeding and reproduction through cooperative asso¬ ciation among cells. The outcome is an amazing diversity of life forms, which the combined efforts of taxonomists, comparative anatomists and physiologists, paleontologists, and, more recently, biochemists and molecular biologists have ordered into a majestic genealogical tree depicting the evolutionary history of extant and extinct animals.1

PHYLOGENY AND ONTOGENY The first elaborate animal tree was drawn by the nineteenth-century German natu¬ ralist and philosopher Ernst Haeckel, an early and enthusiastic disciple of Darwin, as well as an imaginative master in the art of weaving sparse facts into daring, per¬ suasive generalizations. The most famous of these is summed up by the aphorism “ontogeny recapitulates phylogeny,” by which is meant that animals, in the course of their embryological development (ontogeny), go through successive stages that recall the stages of their evolutionary history (phylogeny). Known as the recapitula¬ tion law,2 this statement, though not to be taken literally, expresses a profound truth. Recent acquisitions of molecular biology have shown that development is the main key to animal evolution, which has proceeded largely by way of genetic changes affecting body plans. A possible misapprehension must be corrected. Many of us, when looking at a

188

THE AGE OF MULTICELLULAR ORGANISMS

representation of the animal evolutionary tree, tend to visualize our lineage as pass¬ ing through successive stages in the form of sponges, jellyfishes, worms, mollusks, and so on. This view is false. The animals we are familiar with are terminal twigs on the tiee ol life, final products of long evolutionary histories. Our early ancestors make up the trunk of the evolutionary tree. In order to reconstruct them, we must backtrack mentally from the tips of twigs, through branches of increasing impor¬ tance, down to a major fork where a principal limb separated from the trunk. What we find there are forms considerably less specialized than those that occupy the twigs. By definition, these “fork organisms” make up key ancestral populations that were split by a mutational event into two groups that parted company and started to evolve in different directions. As a rule, these directions were related to two distinct habitats, ot which each gave one of the two groups a reproductive edge over the other. One direction branched out into a complex system of ramifications that pro¬ duced an extant major group of animals. The other direction prolonged the trunk of the tree, up to the next fork from which a new principal limb separated. It is the suc¬ cession of these fork organisms that makes up our ancestry and that we must recon¬ struct. It is an uncertain exercise, since the tree of life is known to us only by the terminal twigs that are still alive and by sparse fossil vestiges whose position on the hee is often difficult to ascertain. However, thanks to comparative sequencing, we are now in a position to evaluate—with still limited but ever-increasing confi¬ dence—the distance that separates two twigs from their last common fork organism.

THE AWAKENING OF ANIMAL LIFE The first successful association experiment on record involved ancient representa¬ tives ot the family choanoflagellates, which are monoflagellated, heterotrophic, aerobic protists so named because their flagellum emerges from the bottom of a food-collecting funnel (khoane in Greek). Such cells may have joined initially into hollow, spherical arrangements combining cooperative propulsion with cooperative feeding.3 With time, further mutations flattened the sphere into a miniature, double-walled pancake, with a back and a belly made of different kinds of cells. The thick, ventral (belly) cell layer served for crawling and food gathering; the thinner, dorsal (back) layer for protection and swimming (see figure 20.1). The animal sometimes raised its central part above the sea floor, creating a space that served as a primitive ali¬ mentary cavity. Like the parent protist, the organism could reproduce sexually undei ceitain conditions, such as excessive crowding. It did so by way of large, nutrient-laden egg cells, which were released after fertilization and developed into copies of the parental organism.

FIGURE 20.1

Some Key Steps in Early Animal Evolution

Ectoderm mi I M i 11 1 1 i-?" I

1

!

i

Endoderm Piploblasts

Anus Ectoderm Mesoderm

dilateral Symmetry

Alimentary Cana Coelom Endoderm Mouth T riploblasts

On top is depicted the evolutionary conversion of a spherical cellular monolayer into a prim¬ itive placozoan, ancestral to all diploblasts, by way of a flattened pouch—with a differenti¬ ated ventral endoderm and dorsal ectoderm—that subsequently bulges in the middle to form an alimentary cavity lined by endodermal cells. Below are shown some early steps in the for¬ mation of triploblasts from diploblasts: (1) development of a third cell layer, the mesoderm, lining an internal cavity, or coelom; (2) body elongation and acquisition of bilateral symme¬ try; and (3) conversion of the alimentary cavity into an alimentary canal, which eventually becomes open at both ends and unidirectional, with a mouth and an anus.

190

THE AGE OF MULTICELLULAR ORGANISMS

This account tells how the basic features of cellular collectivism—association, differentiation, patterning, and a genomically inscribed body plan—could have been first realized in the animal kingdom. The account is imaginary, but its out¬ come is not, being based on a description of Trichoplax adhaerens, a member of the small phylum placozoa, itself a member of the group of diploblasts, which includes all the most ancient known forms of animal life. The term diploblast refers to the two-layered body plan, with one layer, termed ectoderm, derived from the dorsal layer of the ancestral organism and eventually forming the skin of the animals, and the other, named endoderm, derived from the ventral layer of the ancestral organ¬ ism and evolving into a mucosal, digestive lining. Two major lines diverged from the primitive, ancestral diploblast. While one continued to exploit the original body plan, another began to rearrange the two¬ layered pancake into a network of interconnected channels. The cells lining these channels kept water flowing by their beating flagella; they fished out bacteria and smaller food particles from the flowing water, which also provided them with min¬ eral salts and oxygen and carried away waste products. In the meagerly supplied habitat these organisms occupied, this modification proved advantageous. Instead of relying on sluggish creeping to find food, the organisms did better by filtering large amounts of water through their channels. No longer in need of mobility, they became fixed, developed an extensive supporting skeleton, and went on exploiting the advantages of their new body plan by creating an ever more complex labyrinth of cavities and channels. Their present-day descendants are the sponges, whose protein skeleton, cleaned and processed, strokes our skin with a softness no plastic material has yet matched. In the other diploblast line, the tendency of Trichoplax to raise its middle part into a space in which food was retained and digested was accentuated and further exploited. Eventually, the organisms took the shape of miniature double-walled pouches that opened to the outside by a narrow orifice. Thanks to this transforma¬ tion, the organisms gained a segregated digestive cavity, a marked advantage pro¬ vided enough food entered the cavity. Mutations endowing the rim of the pouch with food-catching appendages took care of this requirement. The outcome was a tiny, primitive medusa, the common ancestor of hydrae, polyps, sea anemones, jellyfishes, and other related organisms known jointly as coelenterates. The kinship among these animals—even their animal nature—may not be obvi¬ ous to the naked eye. But look closely and you will see, projecting from myriad tiny chambers in a coral reef or from the bole of a sea anemone, entities built according to the same general plan as a Portuguese man-of-war. A number of species exist that alternate between a fixed polyp form and a free-swimming medusa form. The bodies of all these organisms are constructed in radially symmetrical fash¬ ion around a central alimentary cavity communicating with the outside by a single opening, which serves as both mouth and anus. A variety of tentacles armed with stinging, sometimes deadly poisons, are arranged around the opening, serving to

THE FIRST ANIMALS

191

capture prey and draw it into the cavity where the catch is digested and assimilated. Residues are discharged through the same opening. These animals are built of a number of differentiated cell types, sometimes including muscle and nerve cells. They are equipped to float and drift with the currents. Some move actively by a kind of jet propulsion induced by contraction of the central cavity. They reproduce sexually by means of typical sperm and egg cells. Most are hermaphroditic, but some have distinct male and female forms. Sponges, coelenterates, and a few related animals represent today’s outcome of the diploblast experiment. Their ancestors shared the seas only with algae and sea¬ weeds—and with a multitude of microbes—sometime between 600 and 700 million years ago. If their variety at that time was anything comparable to what it is today, they would have been enough to delight a scuba diver. Visit some tropical, under¬ water “animal garden,” ignore the fish, the crabs and other Crustacea, the worms, the octopuses, the hard-shelled mollusks, and you are left with a view of what the seascape might have been in those early days of animal evolution—and might per¬ haps still be today had not some mutant initiated a new kind of body plan.

THE WORM’S FINEST HOUR Two major changes characterized the new body plan (see figure 20.1). First, the symmetry, from radial, became bilateral; the body shape, from circular, became i elongated; the alimentary pouch became a canal, first blind-ended and later open at both ends with a mouth-to-anus polarization allowing the directional transit and graded digestion of food. With these changes came the emergence of a head, in which nerve cells began to congregate around the mouth to create the first rudi¬ ments of a brain, and the appearance of well-developed excretory and reproductive organs. Preceding, accompanying, or following these developments—no information is available on the time sequence—a third cell layer known as the mesoderm arose from the ventral, or endodermal, layer of the ancestral diploblast, giving rise to the three-layered body plan characteristic of triploblasts, which include the major part of the animal world.4 Still entirely soft-bodied, these newcomers have left no fossil remains. How¬ ever, petrified mud dating back more than 600 million years has kept traces of their trails, tracks, and burrows, revealing signs of their erstwhile importance and diver¬ sity. Their most primitive present-day descendants are the flatworms. Several of these have become adapted to parasitic life and have an atrophic alimentary system, among them the tapeworms, which inhabit the mammalian digestive tract, and the schistosomes, or flukes, the agents of several grave tropical diseases. Next are the nemertine worms, which are the most ancient animals having a one-way alimentary canal, and a number of other primitive, wormlike animals, including the round-

192

THE AGE OF MULTICELLULAR ORGANISMS

worms, or nematodes, which are found everywhere and are said to be the most abundant kind of animal life in the world. Nematodes comprise several parasitic forms that found their niche in mammals, including the more than one-foot-long ascaris worms that thrive in the intestine of horses; the common pinworm familiar to many parents of young children; and the much more dangerous agents of such dreaded diseases as trichinosis, ankylostomiasis (hookworm disease), and filariasis (elephantiasis). The body plan of these lowly worms is understandably described as primitive by zoologists, who have seen so much greater sophistication in insects, fish, birds, and mammals. Yet it would have struck an observer who happened to explore the oceans in those remote days as a marvel of exquisite intricacy. We can get a glimpse of this remarkable complexity by looking at the small nematode Caenorhabditis elegans, currently the most completely known of all animals.5 It consists of exactly 959 cells, each of which has been located with great precision and traced back to its origin from the egg cell through between eight and seventeen successive rounds of mitotic division. Many studies have thrown light on how this development unfolds from the program inscribed in the egg cell’s genome. All sorts of mutations have been induced, single cells have been killed with miniature laser guns, delicate surgical procedures have been practiced, all helping to unravel the network of genetic switches that command the unfolding of the program and to elu¬ cidate the mysterious signals that allow each of the 959 cells to differentiate cor¬ rectly and find its proper place. The emerging picture is mind-boggling in its com¬ plex precision and will require many more years of research before it is completely understood.

THE “MILIEU INTERIEUR” AND THE OXYGEN CONNECTION Roundworms inaugurated, in still imperfect fashion, an evolutionary development of major importance, namely, the opening of an internal cavity, or coelom (from the Greek koilos, hollow). The alimentary cavity, whether with one or two openings, is not a true internal cavity. It communicates with the outside. The coelom does not. In its simplest form, it is a hollow, double-walled sheath completely lined by meso¬ dermal cells (the third cell layer that distinguishes triploblasts from diploblasts) insulating, so to speak, the endodermal alimentary canal from the ectodermal skin (see figure 20.1). In the human body, the coelom is represented mainly by the abdominal and thoracic cavities; its mesodermal lining forms the peritoneum, which surrounds the abdominal viscera, and the pleura, which envelops the lungs. The coelom and the one-way alimentary canal running from mouth to anus were crucial additions to the basic triploblastic body plan. They gave evolution an enor¬ mously expanded range of potentialities to exploit, resulting in a profusion of new

THE FIRST ANIMALS

193

marine animals, some of which, for the first time in history, possessed hard parts. Marked by a sudden abundance and rich variety of fossil remains, many belonging to bizarre, long-extinct animal species, this period is described by paleontologists as the great Cambrian explosion, or radiation.6 (The Cambrian is the geologic period between 600 and 520 million years ago.) What could have triggered this explosion and how does it fit within the geneial framework of animal evolution? A likely factor, though perhaps not the only one, according to the Harvard scientist Andrew H. Knoll,7 was a rise in the oxygen con¬ tent of the atmosphere. The emergence of cyanobacteria equipped with the oxygengenerating photosystem II led, after mineral oxygen sinks became saturated, to a steady rise in atmospheric oxygen content, which, in turn, provoked a majoi ciisis in the prokaryotic world. After oxygen-adapted microbes started appearing, the oxygen produced began to be consumed in increasing amounts, until a steady state was reached in which consumption equaled production and a stable level of atmos¬ pheric oxygen was established. According to Knoll, this level was substantially lower than the present level of 21 percent of atmospheric pressure, and a second important rise occurred in Precambrian times, perhaps as a result of the abundant proliferation of eukaryotic algae. The Cambrian explosion allegedly coincided with this second rise and was made possible by it. Isotopic analysis of Precambrian car¬ bon deposits8 does indeed indicate an excess of photosynthetic activity—which is known to select the lighter carbon isotope 12C against the heavier l3C

over the

capacity of aerobic organisms to oxidize the organic matter made. This disparity would have left a net increase in atmospheric oxygen. As Knoll cautions, this explanation remains hypothetical. However, a relation between the Cambrian explosion and oxygen seems likely. But the cause of the sud¬ den expansion in diversity could have been the increased ability of animals to con¬ sume oxygen rather than the increased availability of oxygen, or it could have been a combination of both factors. All animals have an absolute need for oxygen. This requirement puts severe constraints on marine animals, which must obtain their oxygen from surrounding water, which itself receives it from the atmosphere. Because of the low solubility of oxygen in water, aquatic animals need a large sup¬ ply of freshly oxygenated water and an efficient way of removing oxygen from water. Diploblasts and primitive triploblasts meet these requirements by maintain¬ ing swift water currents along or through their bodies and by having virtually every cell in direct contact with the circulating water. Higher marine animals, however, could not survive without a mechanism for extracting oxygen from the surrounding water and distributing the vital gas to all parts of the body. Evolution had to await the development of such a mechanism before more complex organisms couid arise. Once an effective mechanism was in place, further evolution could have been veiy rapid, thereby producing the Cambrian explosion. Evolution’s key solution to the oxygenation problem was the creation of what the great nineteenth-century French physiologist Claude Bernard9 has called the milieu interieur, an internal fluid of specific composition bathing all the cells of the

194

THE AGE OF MULTICELLULAR ORGANISMS

body. Thanks to this fluid, oxygen taken up by cells in direct contact with sea water could be transferred to more deeply situated cells. Three acquisitions increased the efficiency of this transfer: (1) the formation of gills, specialized surface exchange organs made of thin, highly expanded skin folds allowing a rapid flux of oxygen from surrounding water into the internal fluid; (2) the addition to the internal fluid of special oxygen-carrying molecules—either the red hemoglobin, an iron-contain¬ ing hemoprotein, or the blue hemocyanin, a copper-containing protein—that greatly increased the oxygen capacity of the fluid; and (3) the development of pumps (hearts) to move the fluid and facilitate the transport of oxygen. In their ear¬ liest forms, hearts were no more than contractile thickenings of a tube communicat¬ ing directly with the main body cavity (open circulation). Eventually, tubes sup¬ porting the hearts joined into a closed network (closed circulation) and the milieu interieur became subdivided into two compartments: the circulating blood, present inside the tubes, and the stationary lymph, or intercellular fluid proper. This divi¬ sion made it necessary for the circuit of blood vessels to form two intercalated, highly ramified networks of small, thinly sheathed conduits (capillaries), each net¬ work providing a total surface area large enough for oxygen exchanges to proceed at a sufficient rate. One of these networks traversed the gills and served to collect oxygen from the surrounding water into the blood. The other, traversing the tissues, allowed an efficient delivery of oxygen from the blood to the tissues. Two important fringe benefits accompanied these developments. Nutrients aris¬ ing in the alimentary canal from digested foodstuffs could be collected—this neces¬ sitated a new capillary network around the canal—and distributed throughout the body. Conversely, waste products produced in the body could be discharged into the blood and delivered—through yet another capillary network—to special collec¬ tor cells assembled into excretory organs, the nephridia, or primitive kidneys. The establishment of rudimentary exchange mechanisms for oxygen, nutrients, and waste products by way of a (circulating) milieu interieur marks a turning point in animal evolution. Henceforth, organisms could become more than a few cells deep, and a variety of organs could develop. The cells lining the alimentary canal began to differentiate into several types, some of which formed separate organs (glands) that manufactured digestive enzymes and discharged them into the canal through ducts. Secretion was bom. Mobile cells, specialized in various forms to defend against pathogenic microorganisms, started patrolling the body by way of the internal fluid and, later, by way of the blood. Immunity was initiated. In response to increasing size and bulkiness of the organism and its organs, certain cells turned into builder cells that constructed extracellular supporting frameworks made of proteins and carbohydrate polymers. Contractile cells joined into muscles, appropriately disposed to accomplish coordinated movements. Special reproductive organs, often voluminous, also developed for the formation and maturation of gametes and for ensuring fertilization by a variety of mechanisms, including direct copulation. Finally, to meet the growing need for regulation and coordination, nerve cells wove increasingly complex networks, while modified gland cells mak-

THE FIRST ANIMALS

195

ing chemical transmitters (hormones) discharged their products into the internal milieu instead of into the alimentary canal. Thus was perfected the basic animal blueprint, with its key functions of food capture, digestion, and absorption, oxygen uptake, waste elimination, locomotion, and reproduction, linked by circulation and coordinated by neural networks and chemical transmitters. To an investigative scuba diver, at that time, the world of worms might have appeared as essentially completed, with only details to be added here and there by evolution. What our explorer could not have foreseen is the cre¬ ative power of duplication.

Chapter 21

Animals Fill the Oceans

Musicians have discovered that if they have a good theme, they can create a rich work by repeating the theme many times in different variations. Composers of serial music have exploited this formula to the utmost, by moving from one varia¬ tion to another in almost imperceptible steps. Something similar happened in ani¬ mal evolution. After the basic blueprint was completed, further progress was made by duplication and variation of the blueprint.

BODY DUPLICATION: THE ROAD TO INNOVATION The first step in the new direction was accomplished by a remarkable genetic modi¬ fication that led to the formation of a multisegmented animal looking for all the wodd like a stiing of primitive worms joined end to end by minimal connections. Each segment of this strange creature was, in itself, an essentially complete organ¬ ism, with an alimentary canal, two nephridia, a circulatory network linked to a pair of gills situated in lateral outgrowths of the body, male and female sex organs, a rudimentary innervation radiating from a centralized group of nerve cells (a gan¬ glion), a set of circular and longitudinal muscles, and a surrounding reinforced skin, or cuticle. Separated by incomplete partitions, the segments were linked to each other mainly by the skin; by the alimentary canal, which was continuous; by two large blood vessels, one running along the back, the other along the belly of the oiganism; and by a nerve cord joining the ganglia. In this early form of the organ¬ ism, the head and tail were constructed like the other segments. This organism clearly originated from a number of copies of the same individ¬ ual, linked end to end. Yet a molecular biologist sent to Earth at that time would have found that most genes were present in the organism’s genome in only single copies. Only a small number of genes, all grouped together, were present in as

ANIMALS FILL THE OCEANS

197

many copies as there were segments, arranged along the chromosome in the same order as the segments. The protein products resulting from the translation of these genes were neither enzymes nor structural proteins; they were proteins that reacted with other genes or sets of genes to turn them on or off, indicating that the dupli¬ cated genes belonged to the superfamily of regulatory genes. In the unsegmented ancestral organism, these genes commanded central switches in the realization of the body plan. Duplication of these genetic switches led to repeats of the instruc¬ tions, with formation of repeats of the organism. At first, all segments were identi¬ cal. But soon the duplicated genes underwent different mutations and the segments became different. The first repeated genes to be affected by mutations in this evolu¬ tion were those situated first and last on the chromosome, leading to the develop¬ ment of specialized head and tail parts. Several of the regulatory genes involved in these phenomena have been identi¬ fied and sequenced. They share a highly conserved sequence of 180 base pairs, called the homeobox (from the Greek homos, same). The stretch of sixty amino acids encoded by this box in the corresponding proteins is characteristically shaped to bind to DNA, as befits proteins influencing the transcription of certain genes (transcription factors). Homeotic genes are very ancient. They have been recog¬ nized throughout the animal world and even in plants and fungi.1 Segmentation represents a major mechanism of evolutionary diversification, perhaps the most important one in the history of life. It initiated an extraordinary combinatorial game involving complete, originally viable modules that could be mutated, fused, reduplicated, deleted, and otherwise reshuffled, all by the magic stroke of single or sparse genetic modifications, to offer natural selection a large variety of body plans to test. Investigators working on the fruit fly Drosophila, the central object of classical genetic research, have discovered the amazing creative flexibility of this game. They have been able to produce, by single homeotic gene mutations, headless, two-tailed flies; animals with an extra pair of legs or wings; or strange monsters sporting legs in front of their heads in lieu ol antennae. Using the means of modern molecular biology, these investigators have begun to unravel the remarkable molecular mechanisms whereby homeotic genes control development.

INVERTEBRATES GALORE The first products of this new game were most similar to present-day annelids, so named because their bodies appear like a series of rings (anulus means ring in Latin). The land-adapted, common earthworm is a familiar member of this phylum, which also includes a variety of sea worms, among them a number of organisms (fan worms, peacock worms) that live in solid tubes of their own making and catch prey by means of beautifully colored appendices that they wave fiom the ontice ot their abode. Leeches form another class of annelids.

198

THE AGE OF MULTICELLULAR ORGANISMS

Evolution did not stop at the still fairly repetitious body plans of annelids. After considerable further variations, of which few intermediates are known, it ended up producing the most extensive group of animals living on Earth today: arthropods, or animals with articulated limbs (arthron means joint in Greek and pod comes trom the Greek word for foot), represented in water by the Crustacea (shrimps, crabs, and the like) and the chelicerates (sea spiders and horseshoe crabs) and on land by spiders, scorpions, ticks, centipedes, millipedes, and, foremost, the immense group of insects. Look at a lobster or shrimp. You have no difficulty recognizing the segmented body plan inherited from the ancestral organism. There are one pair of gills and one paii ot legs per segment, but with considerable variation in the specialization of each segment. Several front segments are fused together to form the head, and their legs are converted to a variety of antennae, claws, masticating devices, and other appendages serving as sensory organs or feeding aids. The limbs consist of several articulated parts. The main body and tail segments are largely muscular, while the viscera are grouped in the front part of the body. Circulation is open and the oxygen carrier is hemocyanin. This is in contrast with the more ancient nemertine worms and annelids, whose present-day representatives have a closed circulatory system and use hemoglobin as the oxygen carrier. The arthropod body is entirely covered by a tough carapace made of chitin, a highly resistant, celluloselike carbohydrate polymer. This shell is shed at intervals to permit growth. The molting animal is temporarily vulnerable until it has built a new shell. Amateurs of soft-shelled crabs particularly appreciate this tender state. Somewhere along the annelid-arthropod line, a major branch detached that led to the large phylum of mollusks, which range from the many conches, clams, oys¬ ters, mussels, and sundry other hard-shelled animals we tend to identify with this name to the very different-looking squids and octopuses. The triggering event that started some segmented worm on the way to becoming a mollusk could have been a mutation that endowed a protein constituent of scaly stiuctuies on the back of the animal with the ability to seed the formation of cal¬ cium carbonate crystals. The horny scales became hard, mineralized plates, which gave the animal additional protection and thus provided it and its descendants with a selective advantage. Remnants of this ancestral structure are still visible in chi¬ tons, primitive mollusks that have an elongated structure with bilateral symmetry, an open alimentary canal with a mouth in front and an anus in the rear, two lateral rows of gills, and a series of protective dorsal plates hardened by calcium carbonate deposits. In the further evolution of the ancestral mollusk, the dorsal plates fused into a single shell, segmentation was largely lost, and the body folded and coiled in such a manner that mouth, anus, gills, and excretory and genital outlets all came to be grouped in the front part of the animal, between a head, containing the rudiments of a brain and primitive sense organs sensitive to light (eyes), touch (antennae), and gravitation (otocysts), and a muscular, ventral foot serving for locomotion. Later

ANIMALS FILL THE OCEANS

199

evolution played largely with the shapes of the shells, to the gratification of shell collectors and fossil hunters alike, who both profit from the durability of the min¬ eral deposits. Among the many shapes adopted by mollusks, some may have carried a selec¬ tive advantage, but most variations on the shell theme were probably the result of evolutionary quirks that did not greatly affect the reproductive potential of the ani¬ mals. This fact illustrates an important aspect of evolution: A change does not have to be advantageous to be retained by natural selection, especially under weak com¬ petitive conditions; the change need only not be adverse enough to cause eradica¬ tion. Major developments in the history of mollusks that were presumably the out¬ come of positive selection were the duplication of the shell, leading to the bivalve mollusks, and its atrophy to a mere internal plate in squids, and complete loss in octopuses. This sketchy survey of the vast world of mollusks—the second largest phylum in the animal kingdom, with more than 50,000 living species and almost as many extinct ones—would have completed our description of the animal tree of life—or rather, there would have been no description, for want of a describer—were it not for an astonishing head-to-tail conversion that happened to some ancestral annelid, or, more correctly stated, to its developing embryo.

A FATEFUL FLIP-FLOP: FROM MOUTH FIRST TO MOUTH SECOND In order to understand this new, particularly far-reaching forking of the tree of life, we must take a brief look at embryological development. With Haeckel, who first pointed out the similarity, we note that the developing embryos of animals indeed appear to recapitulate their evolutionary history.2 The cells arising from the early divisions of the fertilized egg first form a sphere, the blastula, which turns into a double-walled pouch, the gastrula, with a single opening, the blastopore. The gastrular cavity later becomes the digestive tract and in all but the most primitive ani¬ mals acquires a second opening and turns into a canal. Here is where development was modified by a mutation to initiate a new evolutionary line. In the animal groups considered so far, the blastopore becomes the mouth and the new opening the anus. They are called protostomes (mouth first) for this reason. The historic flip that started the new line made the blastopore the anus and the new opening the mouth, thereby initiating the deuterostomes (mouth second), the group out of which all vertebrates would someday arise. It is possible that, without this fateful switch, there would be no fish, no amphibians, no reptiles, no biids, no mammals, no humans. Many a biologist in the past has pondered and wondered about the possible mechanism of the astonishingly abrupt change in body plan that caused the

200

THE AGE OF MULTICELLULAR ORGANISMS

deuterostome limb to branch from the protostome trunk. Our wonderment is no less today, but our knowledge of homeotic genes offers a glimpse of a possible explana¬ tion. If a single homeotic gene mutation can replace the head of a fly by a second rear, perhaps one or more such mutations could have interchanged mouth and anus in some ancestral annelid. This is pure speculation, but it is difficult to imagine such a radical change in developmental program occurring otherwise than by some major upheaval of the kind homeotic gene mutations are known to bring about. The fiist consequence of the upheaval detectable in extant organisms (acorn worms) was an anatomical modification that drew together the structures responsi¬ ble for food and oxygen uptake. The front part of the alimentary canal, or pharynx, was converted into a sort of bilateral straining device formed by two opposing rows of narrow slits lined by gills. These branchial slits (branchia means gill in Latin) corresponded to successive segments in the body plan. Animals equipped with this new machinery took in large amounts of water through their mouth and chased the water out through their branchial slits. Oxygen was absorbed from the passing water by the gills, while food particles were retained by the slits, which served as filters, often helped to this effect by fine comblike surface structures. Food col¬ lected by the slits was subsequently sent farther down the alimentary canal. This combined mode of food and oxygen gathering by filtration recalls to some extent the primeval mechanism of maintaining water currents used by sponges, polyps, and jellyfishes.

THE BIRTH OF VERTEBRATES Further key events in the evolution of deuterostomes led to the development of a segmented, hollow structure running along the back of the animal and containing the main parts of the nervous system. If Haeckel’s law is to be believed, an early event in this development was the centralization of nerve cords in a dorsal neural tube, developmental^ derived from an infolding of the ectoderm called the neural crest. Then, underneath the neural tube, there was formed a tough, resilient rod, the notochord, the hallmark, if not always in the adult at least at some embryonic stage, of the whole chordate phylum, of which the earliest representatives are the lancelets. Finally, about 500 million years ago, the neural tube and the notochord became surrounded together by segmented cartilaginous structures, the first verte¬ brae. The most important advantage associated with this development was protection of the h agile neural tube within a solid sheath. Think of the hazards that would beset our highly vulnerable spinal cord were it not protected within our backbone. Seg¬ mentation turned out to be extremely useful in the building of this protective sheath, as the vertebrae could be made of hard, unyielding material without the drawback of an overly rigid body. The segmented spinal column maintained enough flexibility to

ANIMALS FILL THE OCEANS

201

allow all the backbone movements needed for locomotion. The continuous noto¬ chord, which had provided a valuable scaffolding for the construction of the spinal column, later became more of a hindrance than an asset; it eventually came to play only a transient role in early embryological development, subsequently to be broken down. There was a price to pay for the advantage of a segmented backbone, as is well known by many a sufferer of a slipped disk and, more severely, by the paralyzed vic¬ tims of spinal injuries. However, this price did not burden evolution, as it was ex¬ acted only half a billion years later, after some primate adopted an upright posture. The first vertebrates had cartilaginous bones and resembled worms more than fishes, having no jaw and only rudimentary fins. According to the fossil record, some were bizarre', ferocious-looking animals covered with armored plates. Their closest present-day descendants are the lampreys and hagfishes, which are very dif¬ ferent from their remote relatives but share some primitive features with them. The next major development was the formation of a hinged jaw, probably from cartilaginous arches supporting the anterior gill slits. At the same time, the body acquired a variety of fins supported by cartilaginous bones and moved by muscles. The animals turned into powerful swimmers and dangerous predators. Cartilagi¬ nous fish, which include sharks, rays, and skates, are their nearest present-day rela¬ tives. As in the development of mollusks, the last major change was the acquisition of a structural protein capable of seeding the formation of mineral crystals. In the present case, the crystals consisted of mixed calcium phosphate and carbonate, as found in the mineral hydroxyapatite. Resilient cartilaginous structures turned into solid bones. Most present-day fish are descendants of these first bony fish.

THE ECHINODERMS: AN EVOLUTIONARY QUIRK Before the developments that led to the first vertebrates, a bizarre forking event took place in the deuterostome limb shortly after its split from the protostome trunk. As a result, perhaps, of some homeotic gene upheaval, the elongate, bilater¬ ally symmetrical body plan of the ancestral deuterostome, still retained at the larval stage, gave place in the adult to a fivefold symmetrical blueprint in which a greatly compressed and coiled alimentary canal was surrounded by five virtually identical segments. The resulting monster somehow found a favorable niche and thiived, to give rise to the sea urchins, the starfishes, the sand dollars, the sea cucumbers, and other animals characterized by a fivefold radial symmetry. They are grouped undei the name echinoderms (echinos means hedgehog in Greek), even though not all possess the typical spikes of sea urchins. The evolution of invertebrates, from their unicellular ancestor to the first verte¬ brates, is summarized schematically in figure 21.1. As in figure 19.1, the graph

202

THE AGE OF MULTICELLULAR ORGANISMS

shows how key modifications of the body plan of “fork organisms” led to signifi¬ cant evolutionary advances, while the descendants of unmutated organisms give some idea of the body plan of the fork organism, the last ancestor they have in com¬ mon with the more evolved species.

FIGURE 21.1 A Bird’s-Eye View of Invertebrate Evolution

Vertebrates

This figure, similar to figure 19.1, depicts the main steps in the rise of animals in the direc¬ tion of greater complexity, from the ancestral choanoflagellates to the first vertebrates. At each fork, a mutant evolutionary line that underwent the genetic change in body plan indi¬ cated on the left diverges upward from the unmutated line—represented by the arrow curv¬ ing to the right—leading to existing phyla.

Chapter 22

Animals Move Out of the Sea

Once plants and fungi started invading the lands some 400 million years

ago, new pastures became available for animals to exploit. These new opportunities did not long stay neglected. Modifications that had been of no use to aquatic ani¬ mals in earlier days now became advantageous in the changed surroundings. What water-adapted animals mainly needed in order to take advantage of the rich new sources of food offered by land plants was to be able to resist loss of water, to uti¬ lize atmospheric oxygen (respiration), to move on land (for those that lacked such means), and to reproduce away from water. These adaptations were gradual and occurred first on coastal fringes and in marshy areas still exposed to intermittent flooding. Most aquatic animals, except for the lower invertebrates, developed solu¬ tions of one sort or another to the problems of life on land. I shall consider only two types of animals, the arthropods and the vertebrates, which together account for the major part of the terrestrial fauna.

INSECTS AND THEIR RELATIVES: THE GREAT LAND CONQUERORS Arthropods had it easiest, being already shielded by a waterproof covering and equipped with functional legs. However, their fragile gills could not long have resisted desiccation. What helped arthropods to utilize oxygen was the formation of thin, tubular invaginations of their carapace. These tenuous air ducts, or tracheae, progressively developed into a highly ramified network of passages that penetrated all parts of the body, allowing them to be in close contact with outside air. The thin

204

THE AGE OF MULTICELLULAR ORGANISMS

walls of the passages allowed oxygen to diffuse into the tissues and carbon dioxide to diffuse out of them. Body movements served to move air in and out of the ducts, thus restoring the depleted oxygen and removing the accumulating carbon dioxide. All kinds of arthropods developed the same form of respiration by tracheae, including the wormlike, multisegmented centipedes and millipedes; the small crea¬ tures known as wood lice or pill bugs, which are among the rare terrestrial Crus¬ tacea; the spiders and scorpions, which, with other chelicerates, are related to horseshoe crabs; and the innumerable species of insects, almost all of which are ter¬ restrial. These diverse animals did not all inherit tracheae from a common ancestor. What they inherited was a body plan that admitted only one solution to the problem of respiration or, perhaps, favored this solution over others because of some prop¬ erty of the chitinous cuticle that all arthropods have in common. This is a typical instance of convergence. Many marine arthropods reproduce by copulation. Thus, the ancestors of terres¬ trial arthropods did not need water for spermatozoa to find egg cells to fertilize. Their principal requirement in order to reproduce on land was protection and nutri¬ tion of the fertilized egg and the embryo. At first, they simply used water as a medium for larval development, as did their marine forebears and as mosquitoes and many other insects still do today. Then, in the course of time, an amazing num¬ ber of different housings, usually kept moist one way or another, were either adopted or constructed for the successful development of the young away from water. Thanks to popularizers such as Britain’s David Attenborough, television has vividly brought to everyone’s attention the extraordinary operations carried out for the sake of their progeny by dung beetles, termites, bees, wasps, and many other insects. The remarkable coordination and apparent purposefulness of these rituals, faithfully accomplished generation after generation by tiny creatures equipped with a brain no bigger than the head of a pin, have struck many an observer as feats of almost miraculous organization, not readily compatible with a materialistic, Darwinian view of life and evolution. Yet these complex behaviors are child’s play in comparison with the stupendous molecular and cellular events that govern the development of the same animals from fertilized egg cells. Should our eyes be able to follow what happens inside a larva deposited in a beehive alveolus, we would not pay another second’s attention to the construction of the housing itself. Many insects even go through two entirely distinct consecutive developmental programs. From caterpillar to butterfly, from silkworm to moth, from maggot to fly, the animal veritably dies and decomposes within a self-built tomb—cocoon or other pupal covering—leaving alive only some embryonic remnants (imaginal disks). Out of these—the tomb turning into womb—a brand-new organism then arises according to a completely different blueprint. Next to such architectural wiz¬ ardry, inscribed into a couple of feet of DNA, what are a few additional stereotyped gestures serving to build some primitive dwelling? It is like admiring the builders of the Taj Mahal for their ability to make a hut out of straw and mud.

ANIMALS MOVE OUT OF THE SEA

205

AMPHIBIANS: THE FIRST FISH OUT OF WATER Fish also moved out of water, but they had greater obstacles to overcome. They took their time and went through an intermediate, half-aquatic, half-terrestrial stage stable enough to give rise to an important extant class of vertebrates, the amphib¬ ians. We don’t know how the transition occurred but we can hazard some guesses. A key evolutionary event may have been the development, in some fish, of an air-filled pouch communicating with the pharynx. The air in the pouch came from the gills by way of the blood, which circulated through an increasingly rich net¬ work of capillaries surrounding the pouch. An advantage the fish derived from such a pouch was adjustable buoyancy, the main function of what is now the swim blad¬ der. Another advantage was that the fish, in the manner of a scuba diver, carried a reserve of oxygen it could use in case of emergency, when its blood oxygen fell to a dangerously low level. In such an event, oxygen would diffuse in the reverse direc¬ tion, from the pouch into the blood. This adaptation opened the way to breathing, the pouch acting as a primitive lung. We can watch this in our fishbowl when a goldfish surfaces to take a breath of fresh air and, more dramatically, during the dry season in many a tropical lake of Africa, South America, and Australia, where lungfish survive for months in the drying mud, awaiting the next rainy season. This, most likely, is how amphibians “learned” to breathe, while retaining the ability to use dissolved oxygen. Stranded fish capable of breathing would no doubt wriggle their bodies and move their fins in efforts to find shade, moisture, and food. The nimblest at this exercise were animals with two pairs of fleshy, lobed, ventral fins that could help them crawl as well as swim. Fish of this sort were abundant 100 million years ago according to the fossil record. They were believed to be long extinct until a day in December 1938, when one landed in the nets of a fishing trawler off the east coast of South Africa.1 The unusual catch was brought to the attention of the curator of the New London Museum, Marjorie Courtenay-Latimer, who described it to a local ichthyologist, James Leonard Briefly Smith, who recognized it for what it was: a living fossil, described by paleontologists under the name coelacanth. It took four¬ teen years of adventurous episodes, including the posting of rewards in many remote fishing villages along the Indian Ocean and the provision of a special plane by the president of South Africa, Daniel F. Malan, before a second specimen of the rare fish, caught off the Comoro Islands, became available for thorough examina¬ tion. Coelacanths are deep-sea fish and do not use their fins for walking. But they share ancestors with ancient, lobe-finned, freshwater lungfish that invaded swampy lands soon after plants did, some 400 million years ago, thanks to a succession of chance mutations that turned the fins into articulated legs. Such changes would probably not have been retained by natural selection in a watery habitat. On land, they became valuable acquisitions.

206

THE AGE OF MULTICELLULAR ORGANISMS

Before these conquerors could settle definitively in their new surroundings, they had to solve the problem of reproducing on land. Most did not do so and retained the customs of their aquatic ancestors. They spawned in water and their eggs devel¬ oped first into swimming larvae. This was so because water was available every¬ where and there was a lack of selective pressure in favor of true terrestrial repro¬ duction. Evolution rarely moves without some selective inducement. What drove the animals to perfect their breathing and walking machineries was not scarcity of water but abundance of food. This was the time when the great Carboniferous forests began to flourish and to build the surfeit of organic matter that now fuels many of our stoves and furnaces. In addition to plants, plenty of insects, snails, and worms also had become available on land to satisfy animals with more carnivorous tastes. Amphibians thrived in those days but many were later eradicated by the great Permian crisis. Among the survivors, some, such as newts and salamanders, retained the tail of their marine ancestors. Others, among them frogs and toads, kept this appendage only in the free-swimming, fishlike larval stage. The subse¬ quent transformation of a tadpole into an adult frog represents another striking instance of metamorphosis, less dramatic than the total reincarnation exhibited by some insects, but impressive nevertheless. Atrophy of the tail and the sprouting of four legs are among the more conspicuous changes that accompany this transforma¬ tion. These events are triggered by the secretion of thyroxin, an iodine-containing hormone essential to growth in all higher vertebrates. In humans, lack of thyroxin in early developmental stages causes dwarfism and mental retardation. When this substance enters cells, it binds to an intracellular protein receptor, which is thereby turned into an activator of a number of genes. This is an interesting variation on the theme of regulatory supergenes that, like homeotic genes, control by their products the transcription of a number of other genes. There is a twist in this case. The gene product—the thyroxin receptor—is active as a transcription factor only when it has bound the hormone. Other examples of hormones acting in this fashion are ecdysone, which affects molting, pupation, and metamorphosis in insects, and the steroid sex hormones, which control many aspects of the sexual activity of mam¬ mals, including the onset of puberty, the menstrual cycle, and pregnancy. These various phenomena illustrate another fact of general importance: the role of programmed cell death in developmental processes. The decomposition of the caterpillar body and the melting of the tadpole tail are spectacular examples, but there are many others. In the conversion of lobed fins to articulated legs that occurred in the course of the fish-amphibian transformation, the tips of the legs became divided into five fingers, not as a result of budding but through the selec¬ tive death ol intervening tissues. This sculpting is still “recapitulated” in embryonic development. The limbs of a fetus grow first as rounded buds, which are later cut into fingers by selective, programmed cell death.

ANIMALS MOVE OUT OF THE SEA

207

REPTILES “INVENT” THE AMNIOTIC EGG 0

Millennia of sustained drought and wintry cold ushered in the great Permian crisis. The luxuriant Carboniferous forests of ferns and lycopods withered away. The swamps dried out. Marine animals, accustomed to the balmy environment of tropi¬ cal lakes and seas, became extinct in catastrophic numbers. Amphibia also took a heavy toll. But for the fossils unearthed by geologic upheavals, the chance findings of observant wanderers, and the painstaking searches of paleontologists, we would not have the slightest inkling of all this past splendor or of the planetary cataclysm to which it fell victim. As happened many times, life rallied; evolution responded to ecological chal¬ lenges by appropriate adaptions. It even turned disaster into success, driven by the great Permian crisis to accomplish one of its most decisive advances. While seed plants took over the cold, dry swamps left barren by the decimation of sporulating plants, some obscure amphibian suddenly soared into prominence by developing the animal equivalent of the seed: the fluid-filled egg. Instead of delivering fertilized egg cells for development in some body of water—the normal amphibian mode—the female of this key transition species enclosed its fertilized egg cells in a fluid-filled sac, the amnion, within which the embryo could pursue its normal aquatic development. After Claude Bernard’s milieu interieur to bathe all cells and tissues, here was a re-created milieu exterieur to shelter the developing embryo. A hard, porous shell protected this substitute marine incubator, while a highly vascularized membrane, the allantois, produced by the embryo and lining the inner face of the shell, served in gas exchanges and waste disposal. Another sac, filled with a richly nutritious yolk, provided the embryo with necessary foodstuffs. Thus, the complete development of the organism up to a stage where it could survive on land took place within the protective, well-stocked, and appropriately renewed environment of the amniotic fluid. True terrestrial reproduc¬ tion was initiated. The first reptile was bom. This creature enjoyed modest success until the great Permian crisis. After that, reptiles developed and radiated tremendously, to the point of even producing species that lost their legs, while remaining on land, or that returned to an aqueous habitat for living but, paradoxically, moved out of the water, as do sea turtles today, to lay their eggs on land, at great peril to their young. Lizards, snakes, and turtles are the main extant reptiles, but the most spectacular representatives of this group are the dinosaurs, the most celebrated of all fossils. I shall not dwell on the saga of these extraordinary beasts, some of which reached enormous sizes and assumed the most bizarre and, to us at least, terrifying shapes. For a vivid representation of the age of dinosaurs, visit the Peabody Museum at Yale University. A mural, 110 feet by 16 feet, painted by Rudolph Zallinger between the years 1943 and 1947, depicts

208

THE AGE OF MULTICELLULAR ORGANISMS

in arresting colors, amid the contemporary vegetation, the whole history of the dinosaurs, from the emergence of the first amphibia some 400 million years ago down to the great dinosaur extinction, more than 300 million years later.2 We can now see the dinosaurs in full action, thanks to Steven Spielberg’s blockbuster movie Jurassic Park. The disappearance of the dinosaurs 65 million years ago has become one of the most gripping scientific whodunits of all time. Dinosaurs were not the only victims of this mass destruction; they are only the most conspicuous to strike our imagina¬ tions. Many other animals were eradicated at the same time, for example, the beau¬ tiful, spiral-shelled ammonite mollusks. Flowering plants also were decimated, to be replaced for a while by ferns. Many explanations of this mysterious holocaust were proposed, until, in 1978, the American physicist and Nobelist Luis Alvarez, together with his son Walter and other coworkers, made a remarkable observation. They found that a thin layer of sedimentary rocks deposited at the time of the extinction was twenty times richer in the rare element iridium than the adjoining layers.3 Iridium is more abundant in cosmic material than on Earth, and the workers had measured it in order to time the rate at which the material that witnessed the great extinction was deposited at the bottom of the ancient seas. If sedimentation had been fast, the material would have included less cosmic dust, that is, less irid¬ ium. The opposite would be true if sedimentation had been slow. The investigators were looking for modest changes; they had not bargained for the huge increase they found. Here is one more example of serendipity, the magic mother of many a scien¬ tific discovery, a fairy that cannot be courted but sometimes gratuitously favors those who search for truth, even if they do so with the wrong idea in mind. But one must be able to recognize such a blessing. As the great Louis Pasteur once said, chance favors only the prepared mind.4 In the present case, the gift from chance could hardly be missed. The scientists could think of only one explanation for the iridium anomaly: A huge asteroid, six or more miles in diameter, fell on our planet 65 million years ago. First received with considerable skepticism, this suggestion is now widely accepted. Corroborative evidence has been found in many parts of the world, and the probable impact area, almost two hundred miles wide, has been located at a site, Chicxulub, on the north coast of the Yucatan Peninsula in Mexico. How could such an event, a mere prick on the skin of the Earth, cause such a worldwide catastrophe? By sheer brute force. It is estimated that the impact released the equivalent of 100 million megatons of energy, or as much as 10,000 times the energy that would be released by all the atomic bombs of the world exploding at the same time! Clouds of dust, smoke, and soot obscured the sun for years. Raging fires destroyed plant and animal life over large parts of the conti¬ nents. A period of piercing cold (impact winter) was followed by intense warming due to the greenhouse effect of released gases. Acid rain poisoned the waters. Against this doomsday scene, the biblical picture pales to insignificance and the warnings of ecologists become derisory. Once again, however, the irresistible force

ANIMALS MOVE OUT OF THE SEA

209

of evolution came to the rescue and turned disaster into blessing. And what a bless¬ ing, at least from our selfish, anthropocentric point of view, since we might well not be here had an asteroid not hit the Earth 65 million years ago and wiped out the dinosaurs.

THE MAMMALIAN WOMB: THE ULTIMATE GENERATION MACHINE Were the dinosaufs cold-blooded, like all extant reptiles? Or were they warm¬ blooded? This question is the object of a lively debate. Even though we don’t know the answer, we may state that at least one dinosaur branch had or acquired the abil¬ ity to control body temperature around 100°F. These animals remained active in the cold—in contrast to the other, more sluggish reptiles, which attained such a temper¬ ature only when basking in the sun—but they had to pay for this advantage by needing more food. They used their agility to satisfy this requirement and became carnivorous hunters. A thick fur pelt came to cover their bodies, protecting them against heat loss and allowing them to thrive in cold areas where the run-of-the-mill reptile could not survive. Finally, the females adopted the habit, advantageous to the survival and propagation of the species, of covering their eggs until hatched and subsequently shielding their young in a warm clasp. The hungry young, in turn, came to lick the fatty material secreted by skin glands on their mother’s chest. One thing leading to another, in evolution’s usual way of combining chance mutations with natural selection, the secretion turned into milk, and the skin glands into spe¬ cialized, hormonally controlled feeding organs, the mammary glands. Mammals led a modest and inconspicuous existence for nearly 200 million years. They rarely exceeded the size of a rabbit and kept out of the way of the increasingly voracious and ferocious dinosaurs. But when the big test came, the monstrous beasts succumbed, whereas the little, furry animals survived. The rest, as the saying goes, is history. Except that one more development of major importance needs to be mentioned. At some time, a female mammal stopped laying eggs and kept them to incubate and hatch inside her body instead. At first, the young were delivered in a very immature state of development, so fragile as to require immedi¬ ate transfer to a protective ventral skin fold, or marsupium, within which they had access to the mammary glands for feeding. Later, embryos became able to piolong their stay in the mother’s womb and to achieve a much higher degree of develop¬ ment. They succeeded in doing so by drawing nutrients and oxygen from the mater¬ nal blood by means of rootlike extensions, the chorionic villosities, inserted into the wall of the womb, which underwent corresponding adaptations. This intimate fetuswomb connection became the placenta. Today, placentals rule the world, which they have filled with a wide variety of species adapted to every possible kind of environment, including the oceans. Extant

210

THE AGE OF MULTICELLULAR ORGANISMS

egg-laying mammals (monotremes) are rare, for example, the platypus. Marsupials are largely confined to the Australian continent, where they became geographically isolated and never suffered competition from placentals until recent times, when these were brought in by European settlers. If we allow natural selection to have its way, the Australian marsupials will soon be outcompeted by the placentals, as they were in other parts of the world. Among the many branches that grew from the mammalian limb, the tree¬ dwelling primates deserve a special mention. This branch, which detached from the main limb tens of millions of years ago, when dinosaurs still roamed the Earth, went through a long succession of evolutionary forkings and adaptations out of which there emerged, a mere six million years ago, somewhere in East Africa, the twig—indistinguishable from the others at first—through which life entered the Age of the Mind. This will be the subject of the next part of this book. The later steps of animal evolution, from the first vertebrates to the human species, are depicted in figure 22.1, which, like figure 21.1, is drawn so as to high¬ light key genetic modifications of the body plan of “fork organisms.”

THE CONQUEST OF THE SKIES Unlike humans, no animal ever flew because it wanted to fly. The conquest of the skies was entirely a matter of accidental opportunism. The simplest such accident was any anatomical modification that helped an animal to extend the range of a jump by gliding. Flying fish and flying squirrels are examples of animals that are helped in this way by membranous expansions of their fins or limbs. If gliding is useful, natural selection takes care to perpetuate the expansions. A more advanced form consists in the flapping of such expansions as a means of sustaining and pro¬ pelling the body through the air. Flying dinosaurs called pterosaurs are believed to have done just that, with a span that sometimes exceeded thirty feet, and flying mammals, or bats, do it today, using a remarkable sonar device to direct themselves in the dark and locate the insects on which they feed. The most extraordinary and mysterious conquerors of the skies are the insects and the birds. Nobody knows how dragonflies, butterflies, bees, mosquitoes, and other flying insects won their wings. It is not even known whether they inherited their wings from a common ancestor or achieved flying separately by convergent evolution. Unlike the wings of other flying animals, those of insects are not modified limbs. They are formed by flattened outfoldings of the chitinous covering of the animal’s back, which are moved by muscles of extraordinary performance efficiency. How such an amazing arrangement ever came into being is anybody’s guess. The last major bequest of the dinosaurs before they disappeared were the birds. These landed on the world some 150 million years ago, as revealed by the famed Archaeopteryx, a fossil discovered in 1864 in a schist quarry in Eichstatt in Bavaria.

ANIMALS MOVE OUT OF THE SEA

211

FIGURE 22.1

A Bird’s-Eye View of Vertebrate Evolution

Humans

dipedalism, Neoteny 2-

Tree Pwellinp

—

Placenta

-

Womb, Marsupium

Mammary Glands

Amniotic Epps -

Land Adaptation

Ancestral Vertebrates This figure, similar to figures 19.1 and 21.1, depicts the main steps in the rise of animals in the direction of greater complexity, from the first vertebrates to humans. At each fork, a mutant evolutionary line that underwent the genetic change in body plan indicated on the left diverges upward from the unmutated line—represented by the arrow curving to the right— leading to existing phyla.

This weird animal would have passed for a small dinosaur by any test were it not for the imprint of feathers miraculously preserved in the soft stone. Feathers, indeed, turned a reptile into a bird. These remarkable appendages are related to hairs, horns, nails, and scales, and are likewise constructed from a special, tough structural protein, keratin. Feathers obviously did not come into being in one shot, suddenly converting their fortunate owners into flying machines. It must have taken many successive steps to transform a pelt into such a beautiful arrangement of quills and barbs. Flying was out of the question during all that time. It came later,

212

THE AGE OF MULTICELLULAR ORGANISMS

as a fringe benefit, so to speak, albeit one of tremendous value. Some other evolu¬ tionary advantage must have driven natural selection. Many have pondered on the possible nature of this benefit. Better thermal regulation is the explanation consid¬ ered most likely at present. Whether that or another explanation is correct, the phe¬ nomenon itself is a remarkable illustration of the devious ways evolution some¬ times follows to achieve results that have nothing to do with the primary driving force of the process. Once the developing feathers began to allow even the most primitive form of flying, greater proficiency in this extraordinarily advantageous form of transportation became a powerful driving force for the evolutionary improvement of the machinery. Birds, like mammals, have now invaded every pos¬ sible ecological niche and adapted their feeding habits accordingly. Some have even given up their main evolutionary asset and returned to a walking life. According to palynologists, those scientific sleuths who reconstruct the history of the world by looking at fossil pollen, flowering plants enjoyed a remarkable diversification some 50 million years ago. This success is attributed to invasion of the skies by pollen-carrying insects and birds.

THE DRIVING FORCE OF EVOLUTION The history of plants and animals on Earth highlights the groping, unpredictable ways of evolution in its progression toward complexity, mediated at each step by a long-extinct fork organism that offered chance an opening for progress. The two evolutionary lines also illustrate the constraints existing body plans impose upon further advance. Animals have been more “inventive” than plants in this respect, having come up with such spectacular changes as body duplication and the protostome-deuterostome inversion. The evolution of plants has been more conservative, proceeding along the single theme of growth by branching. Each ruled by different selective criteria—plants need light, animals food—the two lines shared a number of problems, to which they evolved comparable solu¬ tions. Increasing body size and complexity was one common problem, which, in both cases, was solved by vascularization. Also shared were the problems posed by the invasion of land, which imposed strict, water-saving measures on both evolu¬ tionary lines. Most important of all common problems was the need for successful reproduction, the selective criterion par excellence. It is significant that both lines adopted sexual reproduction right from the start. This emergency measure of unicellular protists became an essential means of genetic diversification in the evolution of multicellular organisms. Further progress was linked in both lines with the development of more efficient fertilization mecha¬ nisms, better protection of the fertilized egg, and improved fostering of the growing embryo. From aquatic fertilization and development to spores, seeds, and, finally,

ANIMALS MOVE OUT OF THE SEA

213

flowers and fruits in the plant line (see figure 19.1), and to copulation, the amniotic egg, and the mammalian womb in the animal line (see figure 22.1), the trend is unmistakable. Also impressive is the division of reproductive functions between the sexes. In both plants and animals, feeding and sheltering the developing embryo is a female prerogative. The male role is largely restricted to fertilization, compensat¬ ing for a lack of elaborate specializations by an extravagant production of pollen or sperm. The role played by natural catastrophes in the evolution of plants and animals is noteworthy. Evolution is punctuated by massive extinctions, sometimes of cata¬ clysmic proportions. Almost invariably, life’s response has been remarkably inno¬ vative. Apparently, when evolution becomes sluggish, it is not so much for want of an appropriate chance mutation as for the lack of a worthy environmental challenge.

Chapter 23

The Web of Life

In

retracing

the

history of life on Earth (see figure 23.1), we have

looked mostly at the core structure of the tree, the line traced by the successive appearance of living forms of rising complexity. But each major step in this pro¬ gression has also produced side branches that extend their ramifications to the pres¬ ent day. The history of life is not just vertical growth in the direction of complexity; it is also horizontal expansion in the direction of diversity. Each cross section of the tree becomes more varied with advancing time, recapitulating the tree’s previous history by means of what were the terminal twigs of the branches at the time con¬ sidered. A cross section at three billion years ago would show two sturdy, moder¬ ately diversified clusters bearing the archaebacteria and eubacteria existing at that time, almost hiding a tiny, isolated bud, which no observer could have suspected would one day turn into the massive eukaryotic trunk. A cross section at 400 million years ago would show a diversity of bacteria of both types, many kinds of protists, an abundance of algae, some primitive mosses and fungi, a variety of sponges, coelenterates, worms, mollusks, arthropods, and echinoderms, many of them long since extinct, and, sprouting from what we now know to be the main trunk, a number of primitive fish. The display would be richer than the earlier one, yet would include no trees, no flowers, no insects, no amphibians, no reptiles, no birds, no mammals. In such reconstructions, we identify the trunk in retrospect, as the branch that was to lead to the most important innovations in the future. This identification often would not have been evident to contemporaries and is liable to change with time. Today, we place our own species on top of the tree. At least, most of us do so. Ten million years from now, however, we could be on a side branch, or nowhere at all. The new trunk could prolong what looks today as a side branch; it could bear a form of life more complex than the human and beyond the power of our imagination. Cross sections through the tree of life do not just consist of separate dots, as do cross sections through a real tree. In the tree of life, the dots are interconnected by an intricate network of relationships; they form a web. As the dot pattern increased in complexity, so did the web. In this chapter, we shall look at some critical aspects of the development of this web.

THE WEB

OF LIFE

215

THE PRIMORDIAL LINK Most of us think of life as a'creative process, generating form through the synthesis of proteins, nucleic acids, and other specialized molecules. The plastic age has drawn our attention to the importance of biodegradation. Proteins are not intrinsi¬ cally more fragile than many artificial polymers. What renders them fragile is their

FIGURE 23.1

A Summarized History of Life on Earth

Frokaryotes

Eukaryotes

r

A

-^

_ r---—----s

Unicellular Protists

Archaebacteria

Eubacteria

o

0.5

I.o

1.5 IT) 05 Ci

2.0

O sr

O

2.5

cO 3.0

3.5

4.0

4.5

This figure summarizes, in highly schematic form, the history of life on Earth. Notable is the late appearance of multicellular organisms after three billion years of unicellular life. Any horizontal section through this “tree” shows the forms of life present on Earth at the time indicated.

216

THE AGE OF MULTICELLULAR ORGANISMS

biodegradability. Had life not developed the means to break down the products of its own industry, there would be no biosphere, only an inert shell of biopolymers, a ”plastosphere” of the kind human ingenuity is beginning to create. The link between biosynthesis and biodegradation is the primordial link in the web of life. Most likely, it existed already in the first common ancestor of all life. Even if this organism was autotrophic, it must have had the ability to dismantle biopolymers; it must have possessed digestive enzymes. These are the simplest bio¬ logical catalysts and can hardly have failed to arise early. Furthermore, they would have been needed to allow the organism to survive in the absence of an energy sup¬ ply—a phototroph in the dark, for example—by consuming the remains of dead cells or part of its own substance. All autotrophs do this today. For such processes to function, the cells needed safeguards against suicidal self¬ digestion. One such safeguard in present-day bacteria is extracellular discharge of the enzymes concomitantly with their synthesis by ribosomes bound to the cell membrane. Other safety measures involve a variety of chemical controls that keep digestive enzymes inactive inside the cells and unleash them only when and where needed. The main digestive enzymes we secrete into our stomach and intestine are made in this way, as inactive “zymogens,” which are activated only when exposed to the gastric or intestinal milieu. If the enzymes are prematurely activated, as in pancreatitis, deadly damage to the tissues may ensue. A third kind of protection depends on confinement of the enzymes within membrane-bounded digestive pock¬ ets (lysosomes), lined by an enzyme-resistant inner layer. Injuries to this lining may also result in widespread tissue damage, as seen in many pathological conditions. In the beginning, biosynthesis exceeded biodegradation and bacterial life pro¬ gressively covered large surfaces of the Earth with thriving, self-sustaining colonies. These soon turned into feeding layers for mutant forms that had lost the capacity for autotrophic growth and turned into obligatory heterotrophs. As revealed by stromatolites, multilayered associations of autotrophic and heterotrophic bacteria had developed in several parts of the world by 3.5 billion years ago, perhaps earlier. These mats became organized in a manner that reflected the requirements and tolerances of their constituent organisms. The surface was occupied by phototrophs that needed the maximum intensity of light. Below them were phototrophs capable of using the light filtered by the top layers. Then came a number of heterotrophs each adapted to their immediate sur¬ roundings. Individual layers in this colony were connected with those above and below by mutual relationships that tended to stabilize the structure of the colony in a steady state. By necessity, phototrophs adjusted their proliferation to the con¬ sumption ability of the underlying heterotrophs, while the heterotrophs limited their voracity to a level compatible with the maintenance of the food-supplying autotrophs. Thus, stromatolite-generating colonies may be viewed as pseudo-organisms, made of several kinds of cell types and held in dynamic equilibrium by a number of self-regulatory circuits. This organization arose spontaneously and was maintained

THE WEB OF LIFE

217

by automatic mechanisms, guided by no more than the blind screening effect of natural selection, which eliminated the forms that did not fit within the colony’s economy and favored those that did. Comparison of stromatolites of different ages with similar extant colonies suggests that the basic organization of such formations may not have changed substantially over 3.5 billion years. As the tree of life developed, so did the bonds between autotrophs and heterotrophs. These bonds still rule the major equilibria of the biosphere. They have become more complex through the emergence of heterotroph-eating heterotrophs, organisms that feed on autotrophs indirectly, sometimes by way of a long food chain. A squid may owe its sustenance to the biosynthetic activity of phototrophic >»

marine microorganisms (phytoplankton) by feasting on a crab that had eaten the remains of a fish that had enjoyed a meal of phytoplankton-fed shrimp. We may ourselves derive energy from the sun by eating meat that was built in a bull thanks to the presence in the animal’s stomachs of microorganisms able to break down grass into usable nutrients. Dynamic equilibria such as exist within a living stromatolite also stabilize the different parts of the biosphere. A population of foxes cannot outgrow the rabbits on which they feed. In turn, the rabbits are limited in their expansion by the prolif¬ eration of the foxes that kill them. Hence, the numbers of foxes and rabbits go through cyclic oscillations in which the rise of one coincides with the decline of the other. This classic example is known as the Lotka-Volterra cycle, from the names of the scientists who studied it theoretically in the 1920s.1 However, such simple predator-prey interactions hardly depict the complex relationships that link together the components of real ecosystems. Even the simplest of fields or ponds are multi¬ factorial systems stitched into intricate networks by dynamic interactions among the plants, animals, fungi, and microorganisms they contain. Such systems, in turn, join to create larger, more complex fabrics, eventually closing into a single, gigan¬ tic web of formidable complexity that envelops the entire Earth: the biosphere. Understanding this web has become a major aim of ecological research. In spite of extensive field studies and increasingly powerful computer simulations, this area of investigation is still in its infancy, so complex and, to some extent, unpredictable are the interconnections it tries to unravel. A rare insect may control the balance of an entire rain forest because of its role in the pollination of certain essential plants. Some key regulatory principles of general significance remain applicable.

BIOSPHERIC METABOLISM Basically, the biosphere and its constituent subsystems continue to be ruled by the primordial link between autotrophs and heterotrophs, acting as recycling converters of matter and energy. The green mantle traps energy from sunlight, carbon from carbon dioxide, nitrogen from nitrate or atmospheric nitrogen, sulfur, phosphorus,

218

THE AGE OF MULTICELLULAR ORGANISMS

sodium, potassium, calcium, magnesium, iron, and other needed elements from dis¬ solved mineral salts, and water from whatever source is available. These materials are converted into biological constituents, with oxygen as the major byproduct. Part of the store of bio-organic products laid up by the phototrophs enters the food chain and serves, directly or indirectly, to nourish animals and other heterotrophs. These organisms utilize their foodstuffs to build their own constituent molecules and to fuel their energy requirements. In doing so, they use oxygen and break down much of their foodstuffs into carbon dioxide and other waste products. The job is completed by worms, fungi, and bacteria, which decompose dead plants and animals, as well as incompletely metabolized waste products such as uric acid, urea, and ammonia. Thus, the oxygen produced by the phototrophs is consumed, and the carbon dioxide, nitrogen, nitrate, and other mineral constituents they use up are regenerated. Parts of these cycles occur anaerobically, depending on fermenta¬ tion processes or the intervention of electron acceptors other than oxygen. Some autotrophic activities take place without the help of light, fueled by the oxidation of mineral electron donors. By and large, these phenomena are governed by the same kind of self-regulating mechanisms that maintained the primitive bacterial colonies in dynamic equilib¬ rium. Synthesis and breakdown tend to balance each other so that the biosphere is held in a steady state and the main biogenic elements are recycled. But there are local variations and, sometimes, large-scale ones. Peat, coal, and oil deposits remind us of a time when biosynthesis greatly exceeded biodegradation. The fossil record holds repeated examples of major upheavals in the composition of the bio¬ sphere. Not least impressive, as well as worrying because we are responsible, are the increasing threats inflicted upon natural balances by human interventions. In order to understand these problems, we must pay attention not only to the interac¬ tions among living members of ecosystems but also to the interactions between liv¬ ing organisms and their environment.

THE ENVIRONMENT The web of life is intimately connected with the environment by a dense network of mutual interactions. That life depends on the environment is obvious. Temperature, sunlight, rainfall, availability of essential nutrients, and other environmental factors delimit the ability of certain plants to grow in a given area and thereby define the possibility for various animals, fungi, and microorganisms to develop in the area. Less evident, perhaps, are the influences exerted by living organisms upon their environment. Yet these influences were, and still are, of fundamental importance. Without life, our planet would be entirely different from what it is. Life has entirely altered the oxidation-reduction balance of the Earth. The large amounts of ferrous iron that filled the Archaean oceans have become locked up,

THE WEB

OF LIFE

219

partly in the ferric form, in minerals such as magnetite and pyrite. Hydrogen sulfide has been mostly oxidized or trapped in minerals and is now found only in certain volcanic areas, where it emerges from cracks in the Earth and is quickly oxidized by atmospheric oxygen. The'most dramatic change is the rise in atmospheric oxy¬ gen itself, which is essentially due to the activity of phototrophs. This change, in turn, has modified the composition of many rocks and given rise to the ozone layer (the ozone molecule, 03, consists of three oxygen atoms), which now shields the Earth and its inhabitants against excessive ultraviolet radiation from the sun. Another major effect of life is the abundant presence of water on our planet. But for living organisms, the Earth would, like Mars, be almost entirely desiccated today. Its water would have been progressively split by UV irradiation. The result¬ ing hydrogen would have escaped into outer space, and the oxygen would have been trapped by mineral “sinks.” Life splits water also, but in a manner that saves the hydrogen and allows the water to be restored. Living organisms also play an important role in keeping the soil humid and in generating the atmospheric currents that bring rain to the continents. Without life, the land masses would have remained arid and dry, as they largely were until living organisms started migrating out of the oceans. Also largely influenced by life is the manner in which carbon dioxide and its salts, the carbonates, are distributed on the Earth. At present, carbon dioxide consti¬ tutes only 0.0315 percent by volume of atmospheric gases and is dissolved in the oceans at a correspondingly low concentration. Its prebiotic levels may have been as much as one hundred times higher. A predictable consequence of such a high atmospheric content of carbon dioxide would have been a considerable heat reten¬ tion due to what is known as the greenhouse effect (see chapter 30). It so happens, however, that the young sun was cooler in those days and delivered about 25 per¬ cent less heat to the Earth than it does now. The high carbon dioxide content of the atmosphere provided a compensatory blanket, so that the surface temperature of our planet was kept around 20° to 25°C (68° to 77°F), according to the best esti¬ mates by experts.2 As life developed and started using up carbon dioxide, more heat escaped from the Earth, but more also was received from the warming sun, so that the two effects largely canceled each other. The carbon abstracted by life from prebiotic carbon dioxide was woven into the organic fabric of the biosphere and, through Carboniferous luxuriance, partly stored in vast underground deposits from which it is now being returned to the carbon dioxide pool by fossil fuel combustion. A good part of the prebiotic carbon dioxide also became immobilized, mostly as calcium carbonate, in the shells and other structures built by marine organisms. Sedimentation, metamorphism, and resurfac¬ ing by tectonic movements have produced the fossil-studded limestones, marbles, and other calcareous rocks that now emerge in many parts of the world. The white cliffs of Dover, the majestic natural cathedrals carved by hidden rivers running deep below our feet, the gleaming Carrara stones out of which so many master¬ pieces were born would not exist had not life arisen and developed on our planet.

220

THE AGE OF MULTICELLULAR ORGANISMS

The biosphere is thus not just a pellicle of living matter that envelops the Earth like a coat. It is intimately linked to the Earth by myriad reciprocal connections, a giant, sun-powered surface processor that both draws from and acts on the crust, oceans, and atmosphere around it, continually remodeling them and being remod¬ eled by them. Life and Earth are so undissociable that some view them as joined into a sort of planetary superorganism, made of interconnected living and nonliving parts held together by a network of cybernetic relationships. This view has been popularized under the name of Gaia.

GAIA She was the Mother Earth goddess of the ancient Greeks. Long forgotten, except in such words as geology, geography, and geometry, she has recently been revived by James Lovelock,3 a distinguished English scientist, an F.R.S. (Fellow of the Royal Society), the most coveted set of initials to follow a British scientist’s name. A physicist by training, Lovelock is a successful inventor of scientific instruments who enjoys a comfortable income from the patents he earned when he was young. He now lives as the proverbial “gentleman of independent means” in a converted eighteenth-century water mill near the border between Devon and Cornwall, the most southwesterly English counties. Coombe Mill (Experimental Station) is both Lovelock’s home and the computer-crammed laboratory in which he simulates the whims and vagaries of Gaia. What distinguishes Gaia from other global concepts is homeostasis, self-regula¬ tion. In Gaia, life and Earth do not simply interact haphazardly. They do so in a manner that tends to correct the imbalances they inflict on each other. An example investigated theoretically by Lovelock is “Daisyworld.” A planet is seeded with a mixture of dark and light daisies that have the same growth requirements but differ by the proportion of incoming light they absorb and reflect. Dark daisies absorb more light and reflect less than do light daisies. On a cold planet, dark daisies fare better than light ones because they retain more heat. They thus spread progressively from the tropical regions, where they started, to the cooler ones, which they help warm up. If, however, the amount of light falling on the planet increases, as hap¬ pened to the young Earth, the planet may become too hot for daisies to grow. The light daisies, which reflect more light and tend to cool their surroundings, are now favored and outgrow the dark ones. Dark and light daisies thus act as a thermostat. They react to changes in temperature in a manner that opposes the changes; they tend to keep the environmental temperature constant. This simple model recalls the Lotka-Volterra predator-prey model. The difference is that the Daisyworld model involves life and an environmental factor, not two forms of life. Lovelock has progressively refined Daisyworld by introducing daisies of up to twenty different shades and even adding rabbits and foxes. The result is always the

THE WEB

OF LIFE

221

same: self-correction, even in the face of deliberately introduced disturbances. With the help of these experimental simulations, Lovelock has built what he considers an increasingly strong case in support of what he acknowledges was originally just an intuitive hypothesis. According to the Gaia theory, the Earth is a living organism that automatically regulates its environment so as to make it optimal for life. Gaia has been greeted with mixed reactions by the scientific establishment. The concept has been enthusiastically endorsed by Lynn Margulis,4 who has become one of its most ardent proponents. According to the late Lewis Thomas, Lovelock’s observations “may, one day, be recognized as one of the major discontinuities in human thought:”5 The cosmologist Freeman Dyson has adopted the Gaia concept and writes, “Respect for Gaia is the beginning of wisdom.”6 Others, however, have been disturbed by the seemingly teleological character of the concept and by the almost mystical language in which it was worded in the beginning. Lovelock agrees that the style of his early writings may have been mis¬ leading, but vigorously protests that Gaia is a bona fide scientific theory, open to testing by observation and experiment. Ecologists tend to distrust the Gaia concept for a different reason. It depicts the living Earth as a robust organism capable of resisting many insults, not as the fragile structure they see as threatened on all sides by human activities. However, Lovelock hardly deserves to be charged with insensitivity toward environmental causes. He has been critical of what he considers misguided emphasis on certain threats, such as weak carcinogens or nuclear power. But at the same time he has spoken out eloquently against what he calls the three deadly Cs,7 cars, cattle, and chain saws, which he blames for destroying the English countryside. The history of life on Earth offers some support for Lovelock’s general point of view. This history has gone through repeated catastrophes, due to such causes as tectonic movements, volcanic eruptions, climatic changes, and asteroid impacts, that wiped out much of the existing flora and fauna. Each time, life not only rebounded but came up with some decisive innovations. However, it took millions of years for this to happen. We can hardly rely on Gaia’s natural resilience if we wish to save the Earth for our children and grandchildren.

Chapter 2J+

The Virtues of Junk DNA

A major difference between eukaryotes, especially the higher plants and animals, and their distant prokaryotic relatives concerns what may be called their DNA thriftiness. Prokaryotes practice the strictest possible economy in DNA con¬ tent. Their genome contains hardly a single nucleotide that is not involved in cod¬ ing or control. In the words of Harvard chemist and Nobelist Walter Gilbert, the bacterial genome is “streamlined,” probably as a result of strong evolutionary pres¬ sure favoring fast proliferation.1 In striking contrast, the eukaryotic genome is made mostly of noncoding DNA without obvious function, sometimes called “junk” or “ballast” DNA. Less than 5 percent of the human DNA has a coding function. Salamanders do much better—or worse, depending on one’s point of view.2 Some of these animals have twenty times more DNA than we have, with those in the west of the United States beating those in the east by severalfold. Fortunately for our self-esteem, DNA quantity is not by itself a measure of overall quality. Western salamanders are not obviously cleverer than their eastern congeners. Having more DNA does not automatically make sala¬ manders superior to us.

SELFISH DNA The amount of apparently useless DNA in the genome of higher plants and animals requires an explanation. According to Britain’s ethologist Richard Dawkins, the explanation lies in the “selfishness” of DNA.3 The unit of selection is DNA, not the body. The body is no more than a means of replicating DNA, just as a chicken has been said to be an egg’s way of making another egg. To quote Dawkins: “The true ‘purpose’ of DNA is to survive, no more and no less. The simplest way to explain the surplus DNA is to suppose that it is a parasite, or at best a harmless but useless

THE VIRTUES OF JUNK D N A

223

passenger, hitching a ride in the survival machines created by the other DNA.”4 This imaginative concept does not explain the striking difference in DNA economy practiced by prokaryotes, and eukaryotes, nor the fact that, albeit with wide varia¬ tions (remember the salamanders), the proportion of “junk” DNA in the eukaryotic genome tends to increase with increasing evolutionary complexification. Part of the eukaryotic DNA seems to play no evident role. There are “dead” genes, copies of functioning genes that have become useless as a result of some crippling mutation. There are also long linking stretches between genes and large stacks of multiple repeats of the same sequence that have no obvious function. In contrast, the bacterial genome contains no dead genes, no unnecessarily long link¬ ers, no stacks of apparently useless repeats. If, as seems likely, evolutionary “streamlining” is responsible for this continual pursuit of genomic leanness, it appears that eukaryotes were not under the same kind of pressure. Indeed, eukaryotic cells are not continually multiplying as quickly as they can. Their DNA replication is a leisurely affair, which takes only part of the time required by cell division and can be adjusted to any length of DNA simply by repli¬ cating more stretches of DNA simultaneously. This faculty is missing in bacteria, which are limited to a single replication origin. It is thus possible that eukaryotes carry “selfish” DNA from generation to generation because the advantage of get¬ ting rid of it is not sufficient to drive natural selection. On the other hand, the possi¬ bility that this DNA plays a role, for example, in chromosomal structure, or in some other unknown way, cannot be excluded.

SPLIT GENES This is not, by far, the whole story of eukaryotic “junk” DNA, nor even its most intriguing chapter. Noncoding DNA is present not only between genes but also within them. Many genes of eukaryotic organisms consist of discrete segments, numbering from two to more than one hundred. Called exons (because they are expressed), these segments are separated by intervening sequences, or introns, that in most cases are not expressed into anything useful. Exons are short and of rela¬ tively uniform length, more than two-thirds being between 50 and 200 nucleotides long. In contrast, the length of introns is much more variable, ranging from less than 10 to more than 50,000 nucleotides. Split genes are transcribed in toto, exons and introns alike, into correspondingly segmented RNAs. These subsequently undergo an intricate processing such that the introns are excised, generally to be broken down, and the exons spliced into mature RNAs. Imagine interspersing a text with gibberish, printing the whole crazy hodge¬ podge, and then carefully cutting out all the gibberish and pasting together the pieces that make sense. No sane person would willingly add so many apparently

224

THE AGE OF MULTICELLULAR ORGANISMS

unnecessary risks of error to the processing of information. Scrambling genetic texts in this way seems particularly absurd because the gibberish adds greatly to the burden of transcription and replication. Furthermore, not a single letter may be missed or misplaced in splicing, lest the whole message become gibberish itself. Finally, the energy cost of the whole, apparently futile, exercise is far from negligi¬ ble. For all these reasons, no scientist before 1977 would possibly have imagined split genes. “Colinearity” was a dogma. It thus came as an utter surprise, rarely equaled by a scientific discovery, when, in 1977, two molecular biologists, Phillip Sharp, from Boston’s Massachusetts Institute of Technology (MIT), and Britain’s Richard Roberts, working at Long Island’s Cold Spring Harbor Laboratory, inde¬ pendently found unmistakable evidence that a gene was divided into several seg¬ ments that were cut out and spliced together in the RNA transcript.3 This discovery earned its authors a 1993 Nobel Prize. Evolution selected gene splicing and honed it to a remarkable degree of preci¬ sion, which means that split genes conferred substantial advantages commensurate with the risks they entailed. According to Gilbert, the most likely advantage cells gained from split genes was “exon shuffling,”6 the ability to make various combi¬ nations of the same DNA modules and put them to the test of natural selection, the way RNA genes are believed to have been first assembled (see chapter 7). Replac¬ ing the earlier, shorter RNA modules, exons were reshuffled within the genome into a wide variety of different “mosaic” genes. Opportunities for diversity were greatly increased. Cells enjoying this flexibility were saved from becoming progressively constrained within a genomic straightjacket. They kept their options open and retained the capacity to innovate. The same exons have, indeed, been used as building blocks for different genes, thereby allowing certain key peptide motifs to be used repeatedly in different con¬ texts,7 just as the same switches, microchips, and other spare parts are assembled in different ways to make different machines. Exon shuffling is re-enacted in a partic¬ ularly remarkable fashion in each individual in the course of maturation of the immune system (see chapter 14).

THE ORIGIN OF INTRONS At what time in the history of life did introns appear in DNA? This is a hotly debated question. According to the evolutionary record, introns came late and spread slowly through the sole eukaryotic line. Almost nonexistent in prokaryotes, they are rare in lower eukaryotes and tend to increase in number with increasing evolutionary progression. This fact suggests that introns entered the genome during or after the prokaryote-eukaryote transition and subsequently spread like some sort of virus to occupy more and more sites within genes (and also between them, thus

THE VIRTUES OF JUNK D N A

225

accounting for some of the junk DNA). A correlate of this view, though not a neces¬ sary one, is that this invasion of the genome by wandering bits of DNA played a significant role in eukaryotic evolution, by multiplying the number and variety of genetic rearrangements that were offered for natural selection to screen. Surprisingly, a case can also be made for the “antiquity of introns,” as elo¬ quently argued by Gilbert.8 What is known of the modular construction of genes suggests that something akin to exon shuffling, albeit with smaller RNA modules, already played a major role in early protocells. So did RNA splicing, which served an essential function in the early combinatorial game with RNA minigenes, reemerging later in another key capacity to convert RNAs transcribed from split genes into mature RNAs. If the first DNA genes were split by exons, an uninter¬ rupted line of descent might exist between the early form of RNA splicing and its present use in eukaryotes. On the other hand, if exons are a late evolutionary inno¬ vation, one has to explain the revival of RNA splicing after more than two billion years of eclipse. The theory that the first genes were split by introns implies that intron loss acted as a brake to evolutionary progress. Bacteria lost virtually all their ancestral introns and remained prokaryotes to the present day. Lower eukaryotes, such as yeasts, conserved a few introns and evolved further. And so on, up to the highest plants and animals, which have retained the largest number of introns. The idea of innovation depending on prolonging a flexible, unformed state where much is still possible holds an undeniable appeal. It fits with the view that important steps in the further evolutionary modeling of body plans were accomplished by putting off the moment of some definitive developmental commitment. The issue of evolution by gain or loss of introns will not be settled on the basis of theoretical arguments. Facts will decide. In support of the acquisition theory, a num¬ ber of introns have been found to be derived from wandering pieces of DNA, or transposable elements.9 The existence of such elements was discovered in the mid1940s by an obscure American plant geneticist, Barbara McClintock, who concluded from the distribution of variegation patches in corn cobs that these patches must have resulted from the transfer of certain DNA stretches from one daughter cell to another in the course of meiosis. Long ignored, this revolutionary concept was eventually recognized as fundamentally important, earning its modest and retiring author be¬ lated world fame and a 1983 Nobel Prize.10 It is now known that certain segments of DNA in prokaryotes and eukaryotes are equipped with end sequences that allow the segments to be excised from their location and inserted into another site, not only within the same genome but also between cells, between organisms of the same or different species, and even across the prokaryote-eukaryote barrier. Several such intruders have been caught in the act of landing in the midst of a gene and causing a genetic deficiency by inactivating the gene. A number of introns have been identified as transposable elements of clearly recent origin. These findings have not closed the debate, but presently available evidence seems to favor the theory that split genes are a late acquisition.11

226

THE AGE OF MULTICELLULAR ORGANISMS

THE UNIVERSE OF EXONS A growing number of exons present in different genes have been found to be related descendants of common ancestral DNA stretches, suggesting that all the proteins found in nature may have arisen from the combination of a limited number of genetic modules. According to an estimate by the Gilbert school, no more than about 7,000 exons—with a range of 950 to 56,000—could have served in the assembly of all known eukaryotic genes.12 Although this estimate of the “universe of exons” is far from unanimously accepted, the very fact that the problem proved approachable and open to an acceptable solution indicates that the number of differ¬ ent exons with which genes were constructed must have been an extremely minute proportion of the unmanageably immense number allowed by simple statistical estimation—450, or one thousand billion billion billion different possible sequences for a stretch of only fifty nucleotides. This implies that extensive exploration of the combinatorial exon “space” may have been possible even at late stages of plant and animal evolution. The significance of this point was emphasized in chapter 7.

PART VI

THE AGE OF THE MIND

Chapter 25

The Step to Human

Seventy footprints in volcanic ash—two individuals of unequal size walk¬

ing side by side and a third following in the bigger one’s steps—were left to petrify for 3.5 million years in what is now the arid Laetoli area of northern Tanzania until they were uncovered in 1977 by Mary Leakey, of the famous Kenyan family of fos¬ sil hunters.1 These ancient traces bear witness to the existence, in those remote times and in that part of the world, of creatures that walked erect on feet resembling ours. Such creatures must have wandered over much of East Africa at that time. The most famous is a young female named Lucy—after the Beatles song “Lucy in the Sky with Diamonds”—who made the headlines in 1974 when her amazingly complete remains—almost half a skeleton—were found in the Afar region of Ethiopia by Donald Johanson, the founder of the Institute of Human Origins in Berkeley, California.2 Lucy is about the same age as the Laetoli walkers. Her pelvic anatomy indicates that she too walked on two legs. So did the owner of a knee joint, dated 3.9 million years, likewise found by Johanson in the Afar region. These early hominids (prehumans) are now known under the name Australopithecus afarensis. The name Australopithecus, meaning southern (